Structures and Buildings Proceedings of the Institution of Civil Engineers

Volume 168 Issue SB6 Structures and Buildings 168 June 2015 Issue SB6
Pages 421–432 http://dx.doi.org/10.1680/stbu.14.00053
Finite-element analysis of a piled machine Paper 1400053
foundation Received 31/05/2014 Accepted 12/02/2015
Published online 29/04/2015
Fattah, Hamood and Al-Naqdi Keywords: foundations/geotechnical engineering/piles and piling

ICE Publishing: All rights reserved

Finite-element analysis of a
piled machine foundation
Mohammed Y. Fattah PhD CEng Ibtehal A. A. Al-Naqdi MSc CEng
Professor, Building and Construction Engineering Department, University Senior Engineer, Ministry of Education, General Directorate of Educational
of Technology, Baghdad, Iraq Planning, Directorate of School Buildings, Baghdad, Iraq
Mohammed J. Hamood PhD CEng
Assistant Professor, Building and Construction Engineering Department,
University of Technology, Baghdad, Iraq

A case study using finite-element software for the dynamic analysis and structural design of a machine foundation
on piles in homogeneous sandy soil is reported. A parametric study was carried out to investigate the effect of the
foundation geometry, the amplitude and frequency of the dynamic load, and the damping ratio. It is concluded that
as the pile cap thickness increases the oscillation of displacement decreases due to material damping inherent in the
concrete of the pile cap. There is a limit of the pile cap size at which its stiffness governs its dynamic response. Above
this size the weight of the pile cap overrides its stiffness effect, and the additional weight leads to an increase in pile
displacement. When the pile group size increases, the frequency at which maximum displacement occurs increases,
and hence the system becomes more stable against resonance. In the case of changing the pile spacing, the
maximum moment factor IM is always at the pile cap centre, where the load is applied. This factor increases when the
pile spacing increases. The dimensionless displacement factor Iz decreases markedly as the pile cap length increases,
reflecting the increase in displacement with pile cap length.

Notation X horizontal distance from the left side of the foundation

B length of pile cap z distance from the fibre exercising stress y along the
d diameter of pile neutral axis
E Young’s modulus c closed shear transfer coefficient
Es stress–strain modulus of elasticity o open shear transfer coefficient
e mass eccentricity ªc density of concrete
F operating frequency ªs density of soil
F(t ) harmonic dynamic force  coefficient of friction
f c9 uniaxial crushing stress  Poisson’s ratio
fcb ultimate biaxial compressive strength i damping ratio
ft uniaxial cracking stress  ah hydrostatic stress
f1 ultimate compressive strength for a state of biaxial y stress in the y-direction
compression superimposed on  ah z stress in the vertical direction
f2 ultimate compressive strength for a state of uniaxial rst steel ratio
compression superimposed on  ah ø circular operating frequency of the machine
Gb shear modulus below the tip of the pile
Gs shear modulus 1. Introduction
h thickness of pile cap Piles are used extensively for supporting building foundations, for
IM normalised moment machine foundations and for offshore structures. The nature of
Iz normalised vertical displacement the pile response and pile interactions is quite different in all
L pile length three cases. Dynamic loading of piles under buildings may cause
Mx moment in the pile cap in the x-direction large deformations and soil non-linearity. Conversely, machines
mi proportional part of rotating mass may cause only small-amplitude vibrations, and soils may behave
P amplitude of harmonic dynamic load as elastic materials. In offshore structures, the piles are especially
S spacing between piles long and slender, with considerable freestanding lengths.
Sc/c spacing between piles, centre to centre
t time In the case of pile foundations, many efforts have been made to
uz displacement at the centre of the pile cap obtain a match between the observed and predicted response of
W width of pile cap the piles under dynamic loading, and arbitrary modifications to

421
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi

stiffness and damping have been made. However, a better match employed the finite-element method for the geometrical model-
could be obtained when a considerably softened or weakened ling of the continuum, an elasto-plastic constitutive model to
zone was assumed around the piles, simulating disturbance to soil describe the soil behaviour and a step-by-step analysis for
during pile installation. A loss of contact of the soil with the pile numerical simulation. The piles were modelled using three-
for a short length close to the ground surface also improved the dimensional isoparametric finite elements, and the raft was
predicted response (Prakash and Puri, 2010). It has been found modelled using shell elements. A realistic stress–strain behaviour
that analytical models often yield higher natural frequencies and of the soil was simulated using a constitutive model for the soil,
lower resonant amplitudes for the lateral dynamic pile response which consisted of two main yield-surface segments: a pressure-
of single piles than those determined from field tests in horizontal dependent perfectly plastic shear failure surface, and a compres-
vibrations only. This was explained by the fact that the analytical sion cap yield surface. Reul (1998) redefined the structural model
models overestimated the shear modulus and radiation damping by the use of infinite elements at the borders to model the soil as
of the soil. an elastic half-space. Katzenbach (2000) used the same structural
model to carry out parametric studies on the behaviour of piled
2. Behaviour of piles under dynamic loads rafts in Frankfurt clay. They used this geometrical modelling of
The most important sources of dynamic loads are earthquakes the continuum to solve a piled-raft problem, and found that the
and machine foundations or similar. In machine foundations, modelling gave accurate results.
loads are applied on the top of the piles, and the dynamic loads
may be estimated from the imbalances in the machines (Prakash Prakash and Puri (2006) studied the methods of analysis for
and Puri, 1988). The strains in the soils are usually small in determining the response of foundations subjected to harmonic
magnitude. When excited, pile foundations may produce vertical excitation. Analogues based on the elastic half-space solutions
vibrations, combined horizontal sliding and rocking, and torsional were used, and soil stiffness was considered frequency indepen-
vibrations, depending on the nature of the dynamic loads. The dent for the design of machine foundations. They found that the
soil–pile system may be considered elastic, and the relevant soil embedment of a foundation strongly influences its dynamic
properties are the shear modulus Gs and Poisson’s ratio v. If the response.
shear modulus around the pile shaft Gs and that below the tip of
the pile Gb are comparable, the pile behaves like a friction pile. Bhatia (2008) found that the suitability of machine foundations
However, if Gb is much larger than Gs, the pile becomes a depends not only on the forces to which they will be subjected
bearing pile. but also on their behaviour when exposed to dynamic loads,
which depends on the speed of the machine and the natural
Ettouney et al. (1983) presented a semi-analytical solution of the frequency of the foundation. Thus a vibration analysis becomes
dynamic behaviour of pile groups, with some assumptions. It was necessary. Each and every machine foundation does require a
found that the dynamic coupling between the piles in the group is detailed vibration analysis, providing insight into the dynamic
important for vertical, horizontal and rotational group com- behaviour of the foundation and its components, for satisfactory
pliances in the low-frequency range. The effects of the distances performance of the machine. A complete knowledge of the load-
between the piles in the group were presented. It was found that transfer mechanism from the machine to the foundation, and also
the use of frequency-independent correction factors, although a complete knowledge of excitation forces and associated
applicable to the static case, tends to distort the pile-group frequencies are necessary for the correct evaluation of machine
compliances in the low- as well as the high-frequency range. performance.

Ghumman (1985) conducted a comprehensive series of model Livshits (2009) concluded that a modal analysis is required for
tests on penetration testing of piles under vertical vibrations. Both verification of the separation of frequencies. Very strict limits on
the rate of penetration and the total penetration increased with the amplitude of vibrations at machine bearings should be
the dynamic force. This experimentally observed behaviour high- checked in a harmonic forced vibration analysis. The response
lighted the importance of vibrations in inducing the settlement of spectrum analysis provided an estimation of the internal forces
piles. and displacements due to seismic excitation. Structural design of
the turbine generator foundation, made of reinforced concrete,
Liu and Novak (1991) employed the finite-element method to required a series of static analyses on various static and quasi-
examine the behaviour of a raft supported by a single pile at the static loads.
centre. In the analysis, the cap was assumed to be circular and to
contact the soil perfectly. Nine noded isoparametric elements In the present study a numerical solution, using the finite-element
were used to model the cap, the pile and the near-field soil method, was found for the machine foundation problem. The
medium. The cap, the pile and the surrounding soil medium were problem investigates the effect of load frequency by taking a
modelled using finite elements. range of frequencies and dynamic load amplitudes into con-
sideration through studying the response of the pile machine
Katzenbach and Reul (1997) described a structural model that foundation in sandy soil. The pile machine foundation system

422
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi

was modelled as a continuum, including soil, which was not floated piles (i.e. the piles are not driven to a rigid stratum). A
modelled as springs as in previous studies. The study included boundary condition prescribing the vertical displacement with a
the effect of different parameters on the behaviour of the pile value of zero was adopted (i.e. the vertical displacement at the
foundation, including pile diameter, the spacing between piles, lateral and vertical boundary was zero). For the pile tips, the
cap thickness and cap length. Normalised factors are illustrated horizontal displacement was zero at the horizontal boundary but
as curves to give the designer a clear vision through clarifying not zero at the vertical boundary. The analysis process was
the settlement and moments induced in the pile cap. considered at the top central point of the pile cap (node A) at
which the vertical dynamic load was applied. In addition to the
3. Description of the pile machine computation of moments and displacements, the values were
foundation problem normalised along the centre line of the floating pile caps. Interface
In the finite-element solution, a circular cross-section of piles was elements (contact surfaces) were used between the concrete of the
adopted, and the soil was meshed with the eight-node brick pile and its cap and the soil.
elements. Figure 1 shows the pile machine foundation and the
location of the applied load, and Figure 2 shows the models of the The analysis was carried out using the finite-element analysis
pile machine foundation and soil for floating piles. The lateral software ANSYS, which enables engineers to perform the follow-
boundaries of the soil were chosen to be sufficiently far from the ing tasks (ANSYS, 2007)
zone of influence under the vertical load (five times the width of
the pile cap). A lower boundary, which simulates the depth of the j building the computer models or transferring CAD models of
soil layer, was also kept sufficiently far from the pile bottom structures, products, components or systems
(0 .25L, where L is the pile length), as the piles were assumed to be j applying operating loads or other design performance
conditions
j studying physical responses, such as stress levels, temperature
distributions or the impact of electromagnetic fields
Node A j optimising a design early in the development process to
reduce production costs.

The ANSYS program has many finite-element analysis capabil-

ities, ranging from a simple, linear, static analysis to a complex,
y
non-linear, transient dynamic analysis. When a structure is
subjected to cyclic loading, the resulting response is expected to
x be cyclic as well. ANSYS can solve this class of problems
through the harmonic analysis option. The restrictions on a
harmonic analysis are twofold: all loads must be sinusoidal
Figure 1. Finite-element mesh of pile machine foundation, functions of time, and all loads must have the same frequency.
illustrating the location of the applied load Sinusoidal loads are specified through the amplitude, phase angle
and forcing frequency range of the parameter. The amplitude is
the peak value of the load, and the phase angle is the time lag
B ⫽ 2·5 m B ⫽ 3·75 m
s ⫽ 1·5 m
between multiple loads that are out of phase with each other. On
s ⫽ 1·5 m
the complex plane, the phase angle is the angle measured from
W ⫽ 2·5 m

W ⫽ 2·5 m
s ⫽ 1·5 m
s ⫽ 1·5 m

the real axis. Finally, the forcing frequency range is the frequency
range of the harmonic load (number of cycles/time).

The analysis of the pile machine foundation was performed for a

d ⫽ 0·5 m d ⫽ 0·5 m foundation resting on finite isotropic elastic homogenous soil, as
B⫽5m the machines may cause only small amplitudes of vibrations (Liu
s ⫽ 1·5 m
and Novak, 1991; Ottaviani, 1975). The concrete was considered
W ⫽ 2·5 m
s ⫽ 1·5 m

to be linearly elastic (Fleischer and Trombik, 2008; Liu and

Novak, 1991). The value of Poisson’s ratio for most types of soil
are given in Bowles (1997), and values for the static stress–strain
modulus Es for selected soils can also be assumed based on the
d ⫽ 0·5 m
guidelines given by Bowles (1997).
L ⫽ 20 m, h ⫽ 0·5 m
The soil used in the study was homogeneous sandy soil. A typical
Figure 2. Pile cap shapes of floating pile machine foundation range of values of the modulus of elasticity for sandy clay soil is
Es ¼ 25–250 MPa. The range of the Poisson ratio for sandy soil

423
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi

is 0 .2–0 .3. The strains in the soils were assumed to be small in 4. Effect of pile cap thickness
magnitude. Three values for the amplitude were used in the The effect of pile cap thickness was studied by considering three
analysis: 80, 100 and 120 kN. thicknesses (0 .5, 0 .75 and 1 m) for a 2 .5 3 2 .5 m pile cap, with
pile diameter of 0 .5 m, pile length 20 m and a spacing/diameter
The harmonic force in any direction perpendicular to the long- ratio of 3. Figure 3 shows the frequency–settlement curves for the
itudinal rotational axis is floating pile machine foundation under three amplitudes of
vertical dynamic load and for three pile cap thicknesses. It can be
1: F(t) ¼ P sin(øt) seen that the displacement first decreased with an increase in pile
cap thickness, by about 49% for a cap thickness of 0 .75 m, for all
three amplitudes of dynamic load (80, 100 and 120 kN). The
where P is the amplitude of harmonic dynamic load (¼ mie, displacement was increased further in the case of the 1 .0 m pile
where mi is the proportional part of the rotating mass and e is the cap thickness, by about 147% more than the displacement of the
mass eccentricity), t is time, ø ¼ 2F is the circular operating 0 .5 m pile cap, again for all three amplitudes of dynamic load.
frequency of the machine, and F is the operating frequency. The decrease and then increase in vertical displacement for the
three pile cap thicknesses arise because the piles were allowed to
Parameters for the floating pile machine foundation are given in move in the vertical direction. When the thickness was increased
Table 1, and the material properties are given in Table 2. from 0 .5 to 0 .75 m, the stiffness of the pile machine foundation

Case Length of pile Width of pile Thickness of Diameter of Length of pile, Centre-to-centre Damping ratio, i
cap, B: m cap, W: m pile cap, h: m pile, d: m L: m spacing between piles,
Sc/c: m

1 2 .5 2 .5 0 .5 0 .5 20 1 .5 0 .05, 0 .1, 0 .2
2 2 .5 2 .5 0 .75 0 .5 20 1 .5 0 .05, 0 .1, 0 .2
3 2 .5 2 .5 1 .0 0 .5 20 1 .5 0 .05, 0 .1, 0 .2

Table 1. Parameters for the floating pile machine foundation,

with damping

Name Definition Values

pffiffiffiffi
Concrete E Young’s modulus: MPa 4700 f c9 25 742 .96
ªc Density of concrete: kg/m3 2400 2400
ı Poisson’s ratio 0 .15a 0 .15
o Open shear transfer coefficient 0 .2a 0 .2
c Closed shear transfer coefficient 0 .7p a
ffiffiffiffi 0 .7
ft Uniaxial cracking stress: MPa .
0 62 f c9 .
3 395 8
f9c Uniaxial crushing stress: MPa f c9a 30 000
fcb Ultimate biaxial compressive strength: MPa 1 .2f c9 36 000
 ah Hydrostatic stress 1 .157f c9 34 .710
f1 Ultimate compressive strength for a state of biaxial compression superimposed 1 .45f c9 43 .500
on  ah : MPa
f2 Ultimate compressive strength for a state of uniaxial compression superimposed 1 .725f c9 51 .750
on  ah : MPa
Interface  Coefficient of friction 0 .5a 0 .5
Soil Es Young’s modulus: MPa 30 000a 30 000
ªs Density of soil: kg/m3 1800a 1800
 Poisson’s ratio 0 .3a 0 .3
Steel rst Steel ratio 0 .002a 0 .002
a
Assumed values.

Table 2. Material dynamic properties

424
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi

0 P ⫽ 80 kN
0 .75 m, and increased by about 26% for all dynamic loads for
Vertical displacement: mm

⫺0·01 P ⫽ 100 kN
P ⫽ 120 kN pile cap thickness 1 m, with reference to the 0 .5 m pile cap
⫺0·02 thickness case.
⫺0·03
The vertical stress was maximum at frequencies 18, 60 and
⫺0·04
54 rad/s for the 0 .5, 0 .75 and 1 m pile caps for the three
⫺0·05 amplitudes (80, 100 and 120 kN).
⫺0·06
0 6 12 18 24 30 36 42 48 54 60 The vertical stress increased with the amplitude of the vertical
Frequency: rad/s
dynamic load by about 30%, 38% and 25% for the 0 .5, 0 .75 and
Figure 3. Variation of vertical displacement with frequency of 1 m pile caps for amplitude 100 kN and by 50%, 73%, 50% for
load for the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .5 m, amplitude 120 kN, with reference to the 80 kN amplitude case.
B*W ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical
dynamic load 5. Design parameters
A dimensionless factor (the normalised vertical displacement Iz)
was used to express the settlement under the pile cap
was increased, leading to greater material damping. A further
increase in the cap thickness to 1 .0 m increases the self-weight of Pd
the foundation, and hence overrides the stiffness effect, causing Iz ¼
2: uz Es BW
an increase in the displacement. Similar results were found by
Kame et al. (2008) during their work on rafts, which was based
on the classic theory of thick plates resting on a Winkler where P is the amplitude of harmonic dynamic load, uz is the
foundation, and taking into account the effect of the raft displacement at the centre of the pile cap and Es is the stress-
thickness, soil modulus and load pattern. strain modulus of elasticity.

The displacement was found to be at its maximum value at the The normalised moment is given by
same frequency for the three amplitudes of load for the three pile
cap thicknesses: 18 rad/s for 0 .5 m, 42 rad/s for 0 .75 m and
54 rad/s for 1 .0 m. The maximum displacement increased with Mx
3: IM ¼
increasing amplitude of vertical dynamic load by about 25% for PB
100 kN and 50% for 120 kN (compared with the reference case,
80 kN) for all the three pile cap thicknesses.
where IM is the normalised moment per unit length in the pile
Figure 4 illustrates the change in vertical stresses with frequency cap, and Mx is the moment in the pile cap in the x-direction.
for the floating pile machine foundation for three amplitudes of
vertical dynamic load and for three pile cap thicknesses. The The equation takes into account the amplitude of the dynamic
maximum vertical stress decreased by about 66% for all three load as well as the cap dimensions. Mx is calculated as
dynamic loads (80, 100 and 120 kN) for pile cap thickness
h

0
ð2
P ⫽ 80 kN
P ⫽ 100 kN Mx ¼ ( y dzB)z
Vertical displacement: mm

⫺0·005 P ⫽ 120 kN h
4: 2
⫺0·010

⫺0·015
where h is the pile cap thickness, y is the stress along the y-axis
⫺0·020
and z is the distance from the fibre exercising stress y along the
⫺0·025 neutral axis.
⫺0·030
0 6 12 18 24 30 36 42 48 54 60 Figures 5 and 6 indicate the displacement under the centre line of
Frequency: rad/s the pile cap, and the moment generalised over the cap for
Figure 4. Variation of vertical displacement with frequency of different cap thicknesses, respectively. The displacement was at
load for the pile machine foundation (h ¼ 0 .75 m, d ¼ 0 .5 m, the maximum directly under the load for all cap thicknesses; the
B*W ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical settlement was 28% for the 0 .75 m cap thickness and 52% for the
dynamic load 0 .5 m pile cap thickness compared with the reference 1 m cap
thickness.

425
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi

P ⫽ 80 kN
Vertical displacement: mm

0 P ⫽ 100 kN
0

Vertical stress: kPa

⫺200 P ⫽ 120 kN
⫺0·05
⫺400
P ⫽ 80 kN ⫺600
⫺0·10
P ⫽ 100 kN ⫺800
P ⫽ 120 kN ⫺1000
⫺0·15 0 6 12 18 24 30 36 42 48 54 60
0 6 12 18 24 30 36 42 48 54 60 Frequency: rad/s
Frequency: rad/s
Figure 7. Variation of vertical stress with frequency of load for the
Figure 5. Variation of vertical displacement with frequency of pile machine foundation (h ¼ 0 .75 m, d ¼ 0 .5 m,
load for the pile machine foundation (h ¼ 1 m, d ¼ 0 .5 m, B*W ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical
B*W ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical dynamic load
dynamic load

1 .0 m pile diameter increased by about 416% compared with the

0 P ⫽ 80 kN reference case (pile diameter 0 .5 m) due to the increase in self-
P ⫽ 100 kN
⫺500 weight of the pile group.
P ⫽ 120 kN
Vertical stress: kPa

⫺ 1000 For each pile diameter the displacement reached a maximum at

the same frequency for all three amplitudes of load: 18 rad/s for
⫺1500
the 0 .5 m diameter piles, 36 rad/s for the 0 .8 m diameter piles
⫺2000 and 24 rad/s for the 1 m diameter piles. The displacement
increased with increasing amplitude of the vertical dynamic load,
⫺2500 by about 25% for the 100 kN load and 50% for the 120 kN load
0 6 12 18 24 30 36 42 48 54 60 in comparison with the reference 80 kN load.
Frequency: rad/s

Figure 6. Variation of vertical stress with frequency of load for the Figures 4(a), 8(a) and 8(b) depict the change in vertical stresses
pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .5 m, with frequency for the floating pile machine foundation under three
B*W ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical amplitudes of vertical dynamic load (80, 100 and 120 kN) and for
dynamic load three pile cap thicknesses. For the 0 .8 m pile diameter the maxi-
mum vertical stress decreased by about 22%, 25% and 22% under
dynamic load amplitudes 80, 100 and 120 kN, respectively. For the
The highest values of the moment were seen for cap thicknesses 1 m pile diameter the vertical stress decreased by 6%, while for the
0 .5 and 0 .75 m. For the 1 m thick cap the moment was maximum 80 kN amplitude of dynamic load an increase of about 7% and
at X/B ¼ 0 .625, as the stress in the y direction was maximum at 11% was obtained compared with the 0 .5 m pile diameter case.
that point. At a frequency of 60 rad/s the maximum moment
decreased by about 24% and about 82% for the 0 .75 and 1 m The vertical stress was maximum at a frequency of 18 rad/s for
thick caps, respectively, compared with the value for the 0 .5 m the 0 .5 m pile diameter, 36 rad/s for the 0 .8 m pile diameter and
cap. The positive moment refers to hogging while the negative
moment refers to sagging. 500
0
Vertical stress: kPa

As expected, the displacement and the moment in the cap ⫺500

decreased with increasing cap stiffness.
⫺1000
⫺1500
6. Effect of pile diameter P ⫽ 80 kN
The pile diameter directly affects the bearing capacity of the pile. ⫺2000
P ⫽ 100 kN
In this study, three different pile diameters were considered: 0 .5, ⫺2500
P ⫽ 120 kN
0 .8 and 1 .0 m. During the analysis, only the pile diameter was ⫺3000
changed, with the other parameters kept unchanged. Figures 3(a), 0 6 12 18 24 30 36 42 48 54 60
Frequency: rad/s
7(a) and 7(b) show the frequency–settlement curves for the
floating pile machine foundation under three amplitudes of Figure 8. Variation of vertical stress with frequency of load for the
vertical dynamic load (80, 100 and 120 kN) and for three pile pile machine foundation (h ¼ 1 m, d ¼ 0 .5 m,
diameters. The displacement increased with increasing pile B*W ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical
diameter, by about 11% for the 0 .8 m pile diameter under all dynamic load
three amplitudes of dynamic load, while the displacement of

426
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi

Normalised moment per unit length: IM

48 rad/s for the 1 m pile diameter under all three amplitudes.
Thus, the pile diameter increases, the frequency at which the 0·6
h ⫽ 0·5
maximum displacement occurs increases, and hence the system 0·5 h ⫽ 0·75 m
becomes more stable against the resonance condition. 0·4 h ⫽ 1·0 m

0·3
The vertical stress increased with increasing amplitude of the
0·2
vertical dynamic load by about 30%, 25% and 47% for the 0 .5,
0·1
0 .8 and 1 .0 m pile diameters under a load amplitude of 100 kN,
and by 50%, 50% and 79% under a load amplitude of 120 kN, by 0
comparison with the 0 .5 m pile diameter case. ⫺0·1
0 0·125 0·250 0·375 0·500 0·625 0·750 0·875 1·000
X/B
In Figure 9 the dimensionless factor Iz, which expresses the
displacement under the centre line of the pile cap (see Equation Figure 10. Normalized moment along the pile cap, for different
2), is plotted for the three different pile diameters. The maximum thicknesses
displacement of the cap increased by about 2% for the 0 .8 m
diameter pile and by about 12% for the 1 .0 m diameter pile,
compared with the reference 0 .5 m diameter pile. This figure reference value), for all three amplitudes of dynamic load.
indicates the importance of selecting a suitable pile diameter in However, for the 1 .75 m spacing the displacement increased by
order to avoid resonance. about 54% compared with the reference case of 1 .5 m spacing,
for all three amplitudes of dynamic load. For each spacing/
7. Effect of spacing between piles diameter ratio the displacement was maximum at the same
The arrangement of piles can influence the vertical settlement of frequency for all three amplitudes of load: 6 rad/s for 1 .25 m,
the pile cap, the vertical stress and the bending moment general- 18 rad/s for 1 .5 m and 42 rad/s for 1 .75 m spacing. The displace-
ised due to the applied load. The square shape of the pile group ment increased with increasing amplitude of the vertical dynamic
was kept unchanged, while three values of the spacing/diameter load, by about 25% for the 100 kN and 50% for the 120 kN load,
ratio (S ¼ 1 .25, 1 .5 and 1 .75 m) were considered. The diameter compared with the reference load of 80 kN. These results indicate
of the piles was chosen to be 0 .5 m, their length was kept that in the floating pile machine foundation studied, the pile
constant and equal to 20 m, and the cap thickness was 0 .5 m. The spacing ratio (S ¼ 1 .5 m) does not reveal the optimum spacing
behaviour of a pile group having a small spacing between piles from a displacement point of view.
may tend towards that of a block, so to maintain the pile raft
concept the spacing between piles needed to be wide enough to Figures 4(a), 11(a) and 11(b) show the variation in the vertical
allow the raft to participate by taking part of the load and using stresses with frequency for the floating pile machine foundation
the piles strategically as settlement reducers. Figures 3(a), 10(a) for three amplitudes of vertical dynamic load and for the three
and 10(b) show the frequency-displacement curves for the pile spacings. For a spacing of 1 .25 m the maximum vertical
floating pile machine foundation under three amplitudes of stress increased by about 5%, 2% and 4% under dynamic loads
vertical dynamic load (80, 100 and 120 kN) and for three of 80, 100 and 120 kN, respectively, compared with a pile spacing
spacing/diameter ratios of the piles. of 1 .5 m; the difference in stress from the reference case vanished
in the case of 1 .75 m spacing under the 80 kN amplitude load.
The displacement decreased with decreasing pile spacing, by
about 26% less for the 1 .25 m than the 1 .5 m spacing (the 0
Vertical displacement: mm

⫺0·01
⫺0·02
Normalised vertical displacement: Iz

h ⫽ 0·5 ⫺0·03
1000
h ⫽ 0·75 m ⫺0·04
800 P ⫽ 80 kN
h ⫽ 1·0 m ⫺0·05
600 P ⫽ 100 kN
400 ⫺0·06 P ⫽ 120 kN
200 ⫺0·07
0 0 6 12 18 24 30 36 42 48 54 60
Frequency: rad/s
⫺200
0 0·125 0·250 0·375 0·500 0·625 0·750 0·875 1·000
X/B Figure 11. Variation of vertical displacement with frequency of
load for the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .8 m,
Figure 9. Normalized vertical displacement under the pile cap, B*W ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical
with different thicknesses dynamic load

427
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi

The vertical stress was maximum at different frequencies for the with increasing the spacing between the piles. At a loading
three amplitudes of dynamic load. It increased with increasing frequency of 60 rad/s the moment decreased by about 67% for
amplitude of vertical dynamic load, by about 25%, 30% and 20% the 1 .25 m spacing and increased by about 18% for the 1 .75 m
for the pile spacings 1 .25, 1 .5 and 1 .75 m under amplitude spacing compared with the reference spacing of 1 .5 m.
100 kN, and by 48%, 50% and 53% under amplitude 120 kN,
compared with the reference amplitude of 80 kN. As shown in Figure 13, the maximum moment factor IM is always
at the pile cap centre, and increases as the pile spacing increases.
The dimensionless factor used to express the displacement under This is because the span supported by the piles increases with
the pile cap Iz was calculated for the three pile spacings, as pile spacing, leading to greater moments.
shown in Figure 12. As expected, the normalised displacement in
the cap was minimum for the 1 .5 m pile spacing. Figures 12 and 8. Effect of pile cap size
13 show the displacement under the centre line (along the x-axis) Three pile cap lengths were considered in the analysis: the
of the pile cap, and the moment generalised over the cap, reference case of 2 .5 m, and 3 .75 and 5 m. During this analysis
respectively, for different spacing values. The displacement under only the pile cap length was changed, while the other parameters
the load was of maximum value for all spacing cases; the lowest were kept unchanged. Figures 3(a), 14(a) and 14(b) show the
displacement was obtained in the case of the 1 .5 m cap thickness. frequency-displacement curves for the floating pile machine
The displacement increased by 20% for the 1 .25 m and by about foundation for three amplitudes of vertical dynamic load (80, 100
14% for the 1 .75 m pile spacing compared with the reference and 120 kN) and the three different pile cap lengths. The
spacing of 1 .5 m. displacement decreased with increasing pile cap length: the
displacement was 11% less for the 3 .75 m pile cap length,
The normalised moment calculated using Equation 3 increased relative to the reference case, under the three amplitudes of
dynamic load; and the displacement for the 5 m pile cap length
0·05 increased by about 106% relative to the 3 .75 m pile cap length,
Vertical displacement: mm

0 under the same amplitudes of dynamic load. This means that

⫺0·05 there is a limit of pile cap size at which its stiffness governs its
⫺0·10
dynamic response. Above this size, the weight of the cap
overrides its stiffness effect, and the additional weight of the cap
⫺0·15
leads to and increases the displacement of the pile foundation.
⫺0·20 P ⫽ 80 kN
⫺0·25 P ⫽ 100 kN
The displacement was maximum at the same frequency for each
P ⫽ 120 kN
⫺0·30 of the three amplitudes of load for the three pile cap lengths: at
0 6 12 18 24 30 36 42 48 54 60 18 rad/s for the 2 .5 m, 12 rad/s for the 3 .75 m and 60 rad/s for
Frequency: rad/s
the 5 m pile cap length.
Figure 12. Variation of vertical displacement with frequency of
load for the pile machine foundation (h ¼ 0 .5 m, d ¼ 1 m, The displacement was increased with increasing amplitude of the
W*B ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical vertical dynamic load, by about 25% and 50% for the 100 and
dynamic load 120 kN loads, respectively, compared with the 80 kN amplitude.

Figures 4(a), 15(a) and 15(b) show the change in vertical stresses
200
0
with frequency for the floating pile machine foundation under the
⫺200
Vertical stress: kPa

0
⫺400
⫺600 ⫺500
Vertical stress: kPa

⫺800
⫺1000
⫺1000
⫺1200 P ⫽ 80 kN ⫺1500
⫺1400 P ⫽ 80 kN
P ⫽ 100 kN ⫺2000
⫺1600 P ⫽ 100 kN
⫺1800 P ⫽ 120 kN ⫺2500
⫺2000 P ⫽ 120 kN
⫺3000
0 6 12 18 24 30 36 42 48 54 60 0 6 12 18 24 30 36 42 48 54 60
Frequency: rad/s Frequency: rad/s

Figure 13. Variation of vertical stress with frequency of load for Figure 14. Variation of vertical stress with frequency of load for
the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .8 m, the pile machine foundation (h ¼ 0 .5 m, d ¼ 1 m,
W*B ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical W*B ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical
dynamic load dynamic load

428
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi
Normalised vertical displacement: Iz

25 000 0

Vertical displacement: mm
d ⫽0·5 m ⫺0·01
20 000 ⫺0·02
d ⫽ 0·8 m
⫺0·03
15 000 d ⫽ 1·0 m ⫺0·04
10 000 ⫺0·05
⫺0·06
5000 P ⫽ 80 kN
⫺0·07
P ⫽ 100 kN
⫺0·08
0 P ⫽ 120 kN
0 0·125 0·250 0·375 0·500 0·625 0·750 0·875 1·000 ⫺0·09
⫺5000 X/B 0 6 12 18 24 30 36 42 48 54 60
Frequency: rad/s

Figure 15. Normalized vertical displacement under pile cap with Figure 17. Variation of vertical displacement with frequency of
different pile diameters load for the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .5 m,
W*B ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .75 m) due to vertical
dynamic load

three amplitudes of vertical dynamic load and for the three pile
cap lengths. For the pile cap length 3 .75 m the maximum vertical
stress decreased by about 4%, 7% and 4% for dynamic load (along the x-axis) for different pile cap lengths, respectively. The
amplitudes 80, 100 and 120 kN, respectively, compared with the displacement under the load was of maximum value for all cap
reference case. There was no difference from the reference case lengths cases; the case of 3 .75 m cap length recorded the lowest
for the 5 m pile cap length at the 80 kN amplitude load, but for displacement. The displacement decreased by about 6% for the
this pile cap length the stresses decreased by about 3% and 7% at 3 .75 m cap length and increased by 282% for the 5 m pile cap
load amplitudes 100 and 120 kN compared with the 2 .5 m pile length, compared with the reference case in which the cap length
cap (reference) case. is 2 .5 m. This increase is attributed to the additional weight of
the pile cap.
The vertical stress was found maximum at different frequencies
for the three amplitudes of dynamic load. The normalised moment under the load was maximum for cap
lengths 2 .5 and 3 .75 m cases; it was maximum at X/B ¼ 0 .0334 for
The vertical stress increased with increasing amplitude of vertical the 5 m cap length, due to the increase in stresses in the y direction
dynamic load by about 29%, 25% and 25% for the 2 .5, 3 .75 and at that location. The cap moment (Equation 3) increased with
5 m pile cap lengths under the 100 kN load amplitude, and by increasing pile cap length. The maximum moment increased by
50%, 50%, 39% for amplitude 120 kN, compared with the about 9% for the 3 .75 m pile cap length, compared with the
reference 80 kN amplitude load. reference 2 .5 m pile cap length, while for the 5 m cap the maxi-
mum moment increased by about 850%, relative to the reference
The dimensionless factor Iz used to express the displacement case, at frequency of 60 rad/s.
under the pile cap (Equation 2) was calculated. The displacement
in the cap was a minimum for the 3 .75 m pile cap length. Figures It can be concluded that the dimensionless factor Iz decreases
16 and 17 show the displacement factor directly under the centre markedly as the pile cap length increases, reflecting the increase
line of the pile cap, and the moment generalised over the cap in displacement with pile cap length. The increase in the normal-
ised moment IM with pile cap length can be attributed to the
P ⫽ 80 kN
Vertical displacement: mm

0 increase in the unsupported length of the cap.

P ⫽ 100 kN
⫺0·01
P ⫽ 120 kN 9. Comparison between machine
⫺0·02
foundations with and without piles
⫺0·03 Introducing piles obviously alters the displacement of the pile
⫺0·04 cap. With a depth of soil of 20 m below the raft cap, with no
⫺0·05 piles and with the other properties remaining the same, a contact
0 6 12 18 24 30 36 42 48 54 60 surface is created between the raft and the soil. Figures 18–28
Frequency: rad/s show the response in the case of a 2 .5 3 2 .5 m, 0 .5 m thick raft
Figure 16. Variation of vertical displacement with frequency of cap. It can be seen that the maximum displacement decreases
load for of the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .5 m, with the use of piles by about 65% under the three load
W*B ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .25 m) due to vertical amplitudes (80, 100 and 120 kN). The maximum displacement
dynamic load for the machine foundation without piles occurs at a frequency of
30 rad/s.

429
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi

Normalised moment per unit length: Im

0
P ⫽ 80 kN
0.7
⫺500 P ⫽ 100 kN
Vertical stress: KPa

P ⫽ 120 kN
0.6 S ⫽ 1·25 m
⫺1000 0.5 S ⫽ 1·5 m
0.4 S ⫽ 1·75 m
⫺1500 0.3
0.2
⫺2000
0.1
⫺2500 0
0 6 12 18 24 30 36 42 48 54 60 ⫺0.1
Frequency: rad/s 0 0·125 0·250 0·375 0·500 0·625 0·750 0·875 1·000
X/B
Figure 18. Variation of vertical stress with frequency of load for
the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .5 m, Figure 21. Normalized moment along the pile cap, for different
W*B ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .25 m) due to vertical spacing values
dynamic load

0
0
Vertical displacement: kPa

P ⫽ 80 kN ⫺0·01
⫺500 P ⫽ 100 kN
Vertical stress: KPa

P ⫽ 120 kN ⫺0·02
⫺1000 ⫺0·03

⫺1500 ⫺0·04 P ⫽ 80 kN
P ⫽ 100 kN
⫺0·05
⫺2000 P ⫽ 120 kN
⫺0·06
⫺2500 0 6 12 18 24 30 36 42 48 54 60
0 6 12 18 24 30 36 42 48 54 60 Frequency: rad/s
Frequency: rad/s
Figure 22. Variation of vertical displacement with frequency of
Figure 19. Variation of vertical stress with frequency of load for
load for the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .5 m,
the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .5 m,
B*W ¼ 3 .75 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical
W*B ¼ 2 .5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .75 m) due to vertical
dynamic load
dynamic load

0
Normalised vertical displacement: Iz

Vertical displacement: mm

⫺0·02
1000 S ⫽ 1·25 m
800 S ⫽ 1·5 m ⫺0·04
S ⫽ 1·75 m
600 ⫺0·06
400
⫺0·08
200 P ⫽ 80 kN
⫺0·10 P ⫽ 100 kN
0 P ⫽ 120 kN
⫺200 ⫺0·12
0 0·125 0·250 0·375 0·500 0·625 0·750 0·875 1·000 0 6 12 18 24 30 36 42 48 54 60
X/B Frequency: rad/s

Figure 23. Variation of vertical displacement with frequency of

Figure 20. Normalized vertical displacement under pile cap, with
load for the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .5 m,
different spacing values
B*W ¼ 5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical dynamic
load

10. Conclusions
Based on the finite-element analysis of machine foundations frequency for all three amplitudes of load for the three cases
resting on floating piles embedded in homogeneous linear elastic of pile diameters. As the pile diameter increases, the
soil, the following conclusions could be drawn frequency at which the maximum displacement occurred also
increases, and hence the system becomes more stable against
j The displacement reached its maximum value at the same resonance at larger pile diameters.

430
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi

Normalised moment per unit length, IM

P ⫽ 80 kN
P ⫽ 100 kN 5
⫺500
Vertical stress: kPa

P ⫽ 120 kN B ⫽ 2·5 m
4
⫺1000 B ⫽ 3·75
3 B ⫽ 5·0
⫺1500 2

⫺2000 1
0
⫺2500
0 6 12 18 24 30 36 42 48 54 60 ⫺1
Frequency: rad/s 0 0·125 0·250 0·375 0·500 0·625 0·750 0·875 1·000
X/B
Figure 24. Variation of vertical stress with frequency of load for
the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .5 m, Figure 27. Normalized moment along the pile cap, with different
B*W ¼ 3 .75 m*2. 5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical pile cap lengths
dynamic load Vertical displacement: mm

0.03
0 0.02 P ⫽ 80 kN
P ⫽ 100 kN
0.01 P ⫽ 120 kN
⫺500 0
Vertical stress: kPa

⫺0.01
⫺1000 ⫺0.02
⫺0.03
⫺1500 ⫺0.04
P ⫽ 80 kN 0 6 12 18 24 30 36 42 48 54 60
⫺2000 P ⫽ 100 kN Frequency: rad/s
P ⫽ 120 kN
⫺2500 Figure 28. Variation of vertical displacement with frequency of
0 6 12 18 24 30 36 42 48 54 60
Frequency: rad/s load for the machine foundation without piles (h ¼ 0 .5m,
d ¼ 0 .5 m, B*W ¼ 2 .5 m*2 .5 m) due to vertical dynamic load
Figure 25. Variation of vertical stress with frequency of load for
the pile machine foundation (h ¼ 0 .5 m, d ¼ 0 .5 m,
B*W ¼ 5 m*2 .5 m, L ¼ 20 m, S ¼ 1 .5 m) due to vertical dynamic
about 67% for the pile spacing of 1 .25 m and increased by
load
about 18% for the 1 .75 m pile spacing, relative to the reference
case of 1 .5 m pile spacing. This is because the span supported
by the piles increases with increasing pile spacing, leading to
Normalised vertical displacement: Iz

greater moments. The increase in IM with pile cap length can

1000 be attributed to the increase in the unsupported length within
B ⫽ 2·5 m
800 B ⫽ 3·75 m the cap, which leads to an increase in the moment.
B ⫽ 5·0 m j The dimensionless displacement factor Iz decreases markedly
600
as the pile cap length increases, reflecting the increase in
400
displacement with increasing pile cap length.
200
0 REFERENCES

⫺200 ANSYS (2007) ANSYS Manual, Version 11. SAS IP, Inc,
0 0·125 0·250 0·375 0·500 0·625 0·750 0·875 1·000 Canonsburg, PA, USA.
X/B Bhatia KG (2008) Foundations for industrial machines and
Figure 26. Normalized vertical displacement under pile cap, with earthquake effects. ISET Journal of Earthquake Technology
different pile cap lengths 45(1–2): 13–29.
Bowles JE (1997) Foundation Analysis and Design, 5th edn.
McGraw-Hill, New York, NY, USA.
j The results indicate that for the floating pile machine Ettouney MM, Brennan JA and Forte MF (1983) Dynamic
foundation considered the pile spacing ratio (S ¼ 1 .5 m) does behavior of pile groups. Journal of Geotechnical Engineering
not reveal the optimum spacing with regard to displacement. 109(3): 301–317.
j When the spacing between piles is changed, the maximum Fleischer PSt and Trombik PG (2008) Turbogenerator machine
normalised moment factor IM is always at the pile cap centre, foundations subjected to earthquake loading. 14th World
and increases with increasing pile spacing. IM decreased by Conference on Earthquake Engineering, Beijing, China.

431
Structures and Buildings Finite-element analysis of a piled machine
Volume 168 Issue SB6 foundation
Fattah, Hamood and Al-Naqdi

Ghumman MS (1985) Effect of Vertical Vibrations on the Livshits A (2009) Dynamic analysis and structural design of
Penetration Resistance of Piles. PhD Thesis, University of turbine generator. European Built Environment CAE
Roorkee, Roorkee, India. Conference, London, UK, pp. 1–12.
Kame GS, Ukarande SK, Borgaonkar K and Sawant VA (2008) A Ottaviani M (1975) Three-dimensional finite element analysis of
parametric study on raft foundation. 12th International vertically loaded pile groups. Géotechnique 25(2): 159–174.
Conference of International Association for Computer Prakash S and Puri VK (1988) Foundation for Machines, Analysis
Methods and Advances in Geomechanics (IACMAG), Goa, and Design. Wiley, New York, NY, USA.
India, pp. 3077–3085. Prakash S and Puri VK (2006) Foundations for vibrating
Katzenbach R (2000) Piled raft foundation projects in Germany. machines. Journal of Structural Engineering 132(4): 1–39.
In Design Applications of Raft Foundation (Hemsley (ed.)). Prakash S and Puri VK (2010) Foundations for dynamic loads. In
Thomas Telford, London, UK, pp. 323–391. Art of Foundation Engineering Practice (Hussein MH,
Katzenbach R and Reul O (1997) Design and performance of Anderson JB and Camp WM (eds)). American Society of
piled rafts. Proceedings of XIVth ICSMFC97, Hamburg, Civil Engineers, Reston, VA, USA, pp. 517–533.
Germany, vol. 4, pp. 2253–2256. Reul O (1998) Soil–structure interaction of a piled raft foundation
Liu W and Novak M (1991) Soil–pile–cap static interaction of a 121 m high office building in Berlin. Proceedings of 12th
analysis by finite and infinite elements. Canadian European Young Geotechnical Engineers Conference, Tallinn,
Geotechnical Journal 28(771): 771–783. Estonia, pp. 1–12.

WHAT DO YOU THINK?

To discuss this paper, please email up to 500 words to the
editor at journals@ice.org.uk. Your contribution will be
forwarded to the author(s) for a reply and, if considered
appropriate by the editorial panel, will be published as a
discussion in a future issue of the journal.
Proceedings journals rely entirely on contributions sent in
by civil engineering professionals, academics and students.
Papers should be 2000–5000 words long (briefing papers
should be 1000–2000 words long), with adequate illustra-
tions and references. You can submit your paper online via
www.icevirtuallibrary.com/content/journals, where you
will also find detailed author guidelines.

432
Copyright of Structures & Buildings is the property of Thomas Telford Ltd and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright
holder's express written permission. However, users may print, download, or email articles for
individual use.