Ocean Engineering 100 (2015) 35–45

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Dynamic response of offshore jacket platform including foundation

degradation under cyclic loadings
Dongfeng Mao a,b,n, Chao Zhong c, Laibin Zhang a,b, Gui Chu a,b
a
China University of Petroleum-Beijing, Ocean Oil-Gas Research Center, Beijing 102249, China
b
SAWS Key Laboratory For Oil–Gas Safety Engineering Center, Beijing 102249, China
c
COTEC Offshore Engineering Solutions, Beijing, China

art ic l e i nf o a b s t r a c t

Article history: A 1/10 scale model is investigated to study the influence of the foundation degradation on the dynamic
Received 19 August 2014 response of an offshore jacket platform. Experiments were conducted in multi-layer saturated soil under
Accepted 22 March 2015 different water depth. A vibration exciter which amplitude and frequency could be changed was
Available online 10 April 2015
employed to simulate realistic loads. The model was scaled using dimensional analysis according to the
Keywords: artificial mass similarity principle and Buckingham π theorem. The horizontal dynamic response under
Offshore jacket platform the foundation degradation after varying cyclic loadings acted were studied. FE models were designed by
Dynamic response means of the lumped parameter method or the equivalent pile technique to handle the pile-soil
Cyclic loading interaction in ABAQUS to analysis dynamic behavior of the jacket platform. Experimental results shows
Foundation degradation
that the foundation degradation is related to the motion amplitude and the frequency of the piles. The
Experiments
degradation will result in the jacket natural frequency depressed, but it has little impact on the system
Finite element method
damping. Comparison of the numerical and experimental results shows that applying the lumped
parameter model with p-y curve PSI element to simulate soil-pile-upper structure interaction can obtain
much higher accuracy than using the equivalent pile technique. Fair agreement between the experi-
mental and theoretical results was obtained.
& 2015 Elsevier Ltd. All rights reserved.

1. Introduction performed full-scale experiments to validate these values. With a

numerical method, taking factors including wave height, wave period,
The offshore structures industry has been flourishing since 1940s. choice of hydrodynamic coefficients CM and Co, and changes in deck
The continuous increase in water depth for which these structures mass and hysteretic structural damping into consideration, Sunder and
have to be constructed necessitated the use of new modeling Connor (1981) performed a sensitivity analysis on jacket platforms. In
approaches to understand how these structures behave under realistic their calculation an equivalent simplified stick model was used. Some
loads (Wisch, 1998; Adrezin et al., 1996). For an offshore jacket researchers calculated the natural frequencies and mode shapes of
platform, not only huge the vertical weight loads from itself and offshore structures in time domain with modal analysis (Karunakaran
upper models will act on, but also the horizontal cyclic loads from et al., 1997) or used exact Timoshenko pipe elements to determine the
wind, wave, current, ice, earthquake and combination of them will dynamic response of an offshore platform in the frequency domain
continuously act on. The dynamic response of an offshore jacket (Horr and Safi, 2003). Ou et al. (2007) developed a damping isolation
platform supported on piles to marine environment loads has to be system to control the vibration of a steel jacket offshore platform. A 1/
significantly influenced by the characteristics of the foundation 10 scale model of the structure was fitted with the damping isolation
(Asgarian and Lesani, 2009; Mostafa and El Naggar, 2004). Many system and tested on a shaking table with it the ice and earthquake
researchers investigated the dynamic response of offshore jacket loads were simulated. They conducted numerical simulations and
structures. Gudmestad and Moe (1996) recommended a unified compared the numerical results with experimental results, and
approach for the selection of appropriate values for the coefficients simulations and experimental results were in good agreement.
used in the calculation of the hydrodynamic loads when they Elshafey et al. (2009) designed a scale model theoretically and exp
compared the API's and North Sea Design Practice methods and erimentally to investigate the dynamic response of an offshore jacket
platform under random wave loads. The experiments were conducted
n
both in air and in towing tank. Excellent agreement between the
Corresponding author at: College of Mechanical and Transportation Engineer-
experimental and theoretical results was obtained.
ing, China University of Petroleum-Beijing, 18 Fuxue Rd., Changping, Beijing
102249, China. The piles on which the jacket platform is supported move in
E-mail address: maodf@cup.edu.cn (D. Mao). cycles, so as to increase excess pore pressure of soil, weaken the

http://dx.doi.org/10.1016/j.oceaneng.2015.03.012
0029-8018/& 2015 Elsevier Ltd. All rights reserved.
36 D. Mao et al. / Ocean Engineering 100 (2015) 35–45

1-accelerometer 2-electric hydraulic actuator

3-force sensor 4-soil pressure cell,
5-pore water pressure gage 6-soil pond,
7-loading frame 8-jacket model
9-displacment sensor

Fig. 1. Model experiment system. (a) Sketch map and (b) physical map.

Table 1
Dimensions and key physical parameters of model.

No. Parameters Dimensions Similarity factor 1/10 scale mode

1 Physical dimension (L) ½L ¼ ½L CL 0.100

2 Mass density (ρ) ½ρ ¼ ½M½L  3 C ρ ¼ 1:0 1.000
3 Elastic modulus (E) ½E ¼ ½M½L  1 ½T  2 CE ¼ CL 0.100
4 Stress (σ) ½σ ¼ ½M½L  1 ½T  2 Cσ ¼ CL 0.100
5 Vibration frequency (ω) ½ω ¼ ½T  1  Cω ¼ CL
 1=2 3.162
6 Exciting force (F) ½F ¼ ½M½L½T  2 C F ¼ C 3L 0.001
7 Exciting durance (t) ½T ¼ ½T Ct ¼ CL
1=2 0.316
8 Exciting frequency ðω0 Þ ½ω0  ¼ ½T  1 C ω0 ¼ C L
 1=2 3.162
9 Gravitational acceleration (g) ½g ¼ ½L½T  2 C g ¼ 1:0 1.000
10 Displacement (s) ½s ¼ ½L Cs ¼ CL 0.100
11 Velocity ðνÞ ½ν ¼ ½L½T  1 Cν ¼ CL
1=2 0.316
12 Acceleration (a) ½a ¼ ½L½T  2 C a ¼ 1:0 1.000
13 Effective overburden pressure ðσ 0 Þ ½σ 0  ¼ ½M½L  1 ½T  2 C σ0 ¼ C L 0.100
14 Pore water pressure (u) ½u ¼ ½M½L  1 ½T  2 Cu ¼ CL 0.100

Table 2 Table 4
Parameters of sand. Test step.

Soil Water Relative Effective Liquid Plastic Cohesive Angle No. Step
content density unit weight limit limit force c of
w (%) Dr (%) γ (kN/m3) wL (%) wP (%) (kPa) friction 1 Hammer pulse excitation
φ (1) 2 Sine cyclic loading, amplitude A, period T, and no. of cycles N
3 Hammer pulse excitation
Sand 21.1 45 19.6 33 21 0.1 20
Note: loading in displacement control way.

foundation and decrease the lateral soil resistance. There are a

Table 3 quite few publications which deal with the pile–soil interaction (PSI)
Parameters of clay.
and the foundation degradation. Matlock (1970) and Reese et al.
Soil Water Effective Liquid Plastic Cohesive Angle of εc (1974) investigated decay patterns of static resistance under cyclic
content unit weight γ limit limit force c friction φ loads for soft clay, hard clay and sand. The equations of static soil
3
w (%) (kN/m ) wL (%) wP (%) (kPa) (deg) resistance suggested by them were adopted in API standard (API,
2000). And in further, Novak et al. (1978), Nogami and Novak (1980),
Clay 27.9 18.0 51 27 11.6 17.3 0.0217
Nogami and Konagai (1988), El Naggar and Novak (1996) and El
Note: εc is the strain corresponding to half maximum stress in undisturbed soil Naggar and Bentley (2000) investigated the clay or sand dynamic
undrained test. resistances and analyzed the dynamic reactions of a pile under
 D. Mao et al. / Ocean Engineering 100 (2015) 35–45 37

vibration loads. Mostafa and El Naggar (2004) studied the response 2.1. Jacket–piles–soil system
of a fixed offshore platform supported by cluster piles with a finite
element method. They analyzed the effect of the behaviors of soil and Jacket–piles–soil system includes the prototype structure, the
the pile–soil interaction on the dynamic characteristics of a jacket scale piles, and a reinforced concrete soil pond filling with layered
structure supported by cluster piles. They concluded that structure's soils. The prototype structure is a four-legged jacket with a
dynamic behavior is significantly affected by the soil characteristics combination of diagonal and X bracings according to an existing
and PSI. Wang and Feng (2005) studied interactions between offshore jacket platform JZ20-2MUQ. The model jacket and the
liquefying sand and pile using model tests on a shake table. CCS model piles are made of Q235 steel with elastic modulus 200 GPa
(1992) stipulates the equivalent pile technique to simply deal with and Poisson's ratio 0.3. In large scale engineering model test, the
PSI in the way that fixes the lower part of a pile from the point at a artificial mass similarity law and the ignoring gravity similarity
distance of 6–8 times of the pile diameter below mudline. The principle based on Buckingham π theorem have been widely used
calculated nature frequency of an offshore jacket platform under to complement the gravity effect and the inertia effect when
relative longtime cyclic loadings has obvious difference with realistic taking the gravity similarity factor to be 1 (hooping the rigidity of
value because the method do not consider the pile–soil interaction the structure keep constant) (Ling et al., 2004). The artificial mass
and the effect of the foundation degradation on the dynamic should be added on the prototype structure as presented in
responses of a platform sufficiently and comprehensive although following equation shows (Ling et al., 2004):
the method can be used to estimate the behaviors of offshore
m0 ¼ c3L mP  mm ð1Þ
structures quickly (Shang et al., 2008).
This paper designed an experimental system including a scale where m0 is the artificial mass, mP is the initial mass of the structure,
model with superstructure, piles and soil to investigate the influence mm is the mass of the prototype structure. c3L mP is the total mass of the
of foundation degradation to dynamic lateral characteristics of an model because the mass similarity factor is c3L , cL is the length
offshore jacket platform. The platform horizontal dynamic response similarity factor. Since length, time and mass are chosen as basic
in condition of the foundation degradation after varying cyclic dimensions, the dimensions and key physical parameters of the model
loadings acted were studied. Soil pressure cells and pore-water are obtained as shown in Table 1 (Li et al., 2012). The model jacket is
pressure gages were used to estimate the reaction of the foundation totally 6 m high. The piles were constructed using pipes with external
and the soil pore pressure, respectively. Experimental results were diameter 90 mm, thickness 3 mm and length 3 m. The whole wei-
compared with the results obtained from the finite element models ght of the structure is 511 kg. For the pond, its length  width 
which were designed in ABAQUS by means of the lumped parameter height¼ 5.0 m  5.0 m  3.5 m. The layered soils inside the pond from
method or the equivalent pile method to validate the finite element the top to the bottom are clay, sand and clay with same thickness 1 m,
models. The finite element models can be used to carry out a large respectively. The soil used in the experiment had been naturally
number of simulations in a practical and less expensive way for
studying the dynamic behavior of offshore structures, especially in
case of the foundation degradation after cyclic loading acted.

2. Experimental setup

The experiment system had been developed as shown in Fig. 1.

It consists of three parts: a jacket–piles–soil system, an electric
servo loading system, and a sensor and data acquisition system.

x
Loading
y

Fig. 3. Single pile scheme.

Fig. 2. FE model of jacket–pile–soil. Fig. 4. The p–y curve at different depths below mudline.
38 D. Mao et al. / Ocean Engineering 100 (2015) 35–45

Fig. 5. The first three orders of model shapes. (a) mode shape#1 bending about x (b) mode shape #2 bending about y and (c) mode shape #3 torsion about z.

consolidated for more than 4 months after multi backfilling and

water-flooding. The soil parameters were obtained by the soil test as
shown in Tables 2 and 3.

2.2. The electric servo loading system

A high-precision electric servo loading system that mainly consists

of an electric servo hydraulic actuator, a driving controller, a computer
and controlling software is set up to simulate the marine environ-
mental loads. The stroke, the maximum velocity and the maximum
load of the hydraulic rams are 0.35 m, 0.2 m/s and 50 kN, respectively.
The controller combing with a force sensor and a displacement sensor
could control the magnitudes of the displacement and the force of the Fig. 6. Comparison of bending moment.
hydraulic ram. Various cyclic loadings with frequencies from 0 Hz to
5 Hz and pulse loadings could be conducted by developed loading
programs installed in the computer.
Table 5
2.3. Sensor arrangement and data acquisition system Natural frequencies and mode shapes of the first three orders of model jacket in
deferent water depth.
As shown in Fig. 1(a), there were 15 groups of half bridge meas-
Order Frequency (Hz) Mode shape Relative deference
urement strain gages mounted on the piles and 15 miniature soil # (%)
pressure cells laid very close to the piles in the soil uniformly in Water depth Water depth
longitudinal with interval 20 cm. The strain gages are adopted to 0m 1.5 m
measure the bending moment of the piles. The miniature soil pressure
1 7.63 7.61 Bending 0.26
cells are employed to test the soil resistance directly to obtain the p–y
about x
curves in different depths. The miniature pore water pressure gages 2 7.95 7.97 Bending  0.25
were placed in soil just before and after the piles in the direction of about y
loading. Two accelerometers were mounted at each main pipe joint in 3 9.55 9.51 Torsion about 0.42
z
two horizontal directions on the upper part of the model jacket. A
displacement sensor was connected to the pile at mudline position to
measure the displacement there. The pulse responses of 36 main pipe
joints could be recorded and the natural frequencies would be given
with a frequency response analysis module of MaCras software. Table 6
Numerical simulation results and experimental results of the first three orders
2.4. Loading (water depth 0 m).

Order Frequency (Hz) p–y curve (equivalent pile)

The sine cyclic displacement loads were exerted on a loading # relative deference to experiment
frame welded to the model jacket at a position of 1.8 m above Experiment p–y Equivalent (%)
mudline (ice load area). The displacement loading amplitude A curve pile
varies from 50 mm to 200 mm with step growth 50 mm and each
1 7.63 8.17 9.31 7.1 (22.0)
amplitude could match different periods T such as 10.0 s, 5.0 s,
2 7.95 8.57 10.25 7.8 (29.0)
2.0 s, 1.0 s and 0.5 s, respectively. Number of cycles N is 200. Before 3 9.55 10.05 11.75 5.2 (23.0)
and after cyclic loading acting, a pulse exciting was acted and its
 D. Mao et al. / Ocean Engineering 100 (2015) 35–45 39

Fig. 7. Changes of excess pore pressure. (a) A¼ 100 mm, T ¼ 10 s cyclic loading and (b) A ¼100 mm, T ¼ 0.5 s cyclic loading.

response curve was recorded in order to analyze impact of the soil

degradation by comparing the change between them. Table 4
demonstrates one cycle of the test steps. In test keeping the
magnitude of the pulse exciting constant is very important.

3. Finite element models

The dynamic response of the model jacket platform is

described by a second-order matrix deferential equation as follow-
ing:
€ þC xðtÞ
M xðtÞ _ þ KxðtÞ ¼ fðtÞ ð2Þ
Where M is mass matrix, C is damping matrix, K is stiffness matrix,
€
xðtÞ; _ and xðtÞ are acceleration, velocity and displacement vec-
xðtÞ
tors, respectively, and fðtÞis excitation.
The ABAQUS software is used to solve Eq. (2). The main
members of the scale jacket were modeled using BEAM element
(Fig. 2). In general, the soil stiffness is highly nonlinear for PSI
problem because soil has not only the viscous damping but also
the radiation damping. There are three ways could be used to deal
with PSI: the lumped parameter model, the equivalent pile
technique, and contact element method (Dongfeng et al., 2015).
In present paper, former two ways were used. Fig. 8. Lateral displacement distribution of pile.

3.1. Equivalent pile technique which made of same material and same dimensions as the scale jacket
piles that were conducted to be tested to obtain static p–y curves in
According to CCS (1992) , in calculation, the damping and the same soil pond. The saturated soil p–y curves that yielded by
stiffness matrices are assumed to be linear, and the low part of the carrying out six order polynomial fitting (Wang and Feng, 2005) with
equivalent pile is fixed in a rigid foundation. The length of the the average values of three groups of experimental pile bending
equivalent pile is determined by the way of keeping the rigid moment data are demonstrated in Fig. 4.
behavior of the equivalent pile same as that of the pile–soil
system. The lower part of the equivalent pile from the point at a 3.3. Numerical simulation result
distance about 6–8 times of the pile diameter below mudline is
fixed. In this paper 6 times of the pile diameter was used. For the purpose of comparing the simulation with the experi-
mental results, the linear perturbation step was used to vibration
3.2. Lumped parameter model mode and frequency extraction. The loads used in the numerical
simulation should be same with the loads in the experiment. With
In essence, the lumped parameter model is to institute the the CCS equivalent pile method, the first twenty order mode
foundation dynamic stiffness with an equivalent mass–spring–damp- shapes were calculated. The first three order modes are bending
ing system. The scale piles could be simulated with a string of discrete about x and y-axes, and torsion about z-axis as shown in Fig. 5. It
particles or a discrete continuous beam. The foundation which is far should be noted that the x-direction is taken parallel to the long
away from the structure is simplified as a semi-infinite rigidity body. side of the model, y-direction is taken parallel to the short side and
In present paper the discrete continuous beam is applied. Winkler z-axis is taken perpendicular to x-y plane.
model was adopted to simulate the equivalent soil since it ignores the
soil shear, therefore it could be used to handle layered and nonlinear
soil through placing the soil reaction with nonlinear springs. Here p–y 4. Experiment results
curves in different soil depths were taken to define the pipe–soil
constitutive relation to form PSI elements by input key parameters of 4.1. The bending moment of the single model pile
‘PIPE-SOIL STIFFNESS’ including linear and nonlinear parts in ABAQUS.
For four piles, there are 76 PSI elements set up totally along the The bending moment distributions of the single pile were calcu-
loading direction as shown in Fig. 2. Fig. 3 shows a single model pile lated with p–y curve PSI element and tested under three differ
40 D. Mao et al. / Ocean Engineering 100 (2015) 35–45

ent loadings. Their results are shown in Fig. 6. Excellent agreement natural frequencies are shown in Table 5. The experimental results
between the numerical and the experimental results shows the indicate that the water depth has little effect on the natural
validation of the finite element method with p–y curve PSI element frequency of the offshore structure.
in present paper. The comparisons of the numerical simulation and the experimental
results are listed in Table 6. The relative differences between the
4.2. The natural frequency of the model jacket numerical simulation results with the p–y curve PSI element method
and the experimental results locate from 5.2% to 7.8% whereas the
The natural frequencies and mode shapes of the prototype value between the numerical simulation results with the equivalent
jacket in different water depths were obtained and the first three pile method and the experimental results lies from 22% to 29%.

Soil resistance/(kN/m) Soil resistance/(kN/m)

-10 0 10 20 -10 0 10 20
0.0 0.0

-0.5 -0.5

-1.0 -1.0

Depth/m
A=50 mm, T=10 s A=100 mm, T=10 s
Depth/m

-1.5 -1.5
Before loading
Before loading
-2.0 After loading -2.0
After loading

-2.5 -2.5

-3.0
Soil resistance/(kN/m) Soil resistance/(kN/m)
-10 0 10 20 -10 0 10 20
0.0
0. 0

-0.5 -0.5

-1.0 A=150 mm, T=10 s -1.0

A=200 mm, T=10 s
Depth/m

Depth/m

-1.5 -1.5
Before loading Before loading
-2.0 After loading -2.0 After loading

-2.5 -2.5

-3.0 -3.0

Fig. 9. Comparisons of soil resistances before and after cyclic loadings.

Soil resistance/kN/m

20
A=100mm, T=10s A=150m, T=10s
A=50mm, T=10s A=200mm, T=10s
15

10

0
-25 -20 -15 -10 -5
-5 0 5 10 15 20 25

-10
Displacement/mm

-15

-20

Fig. 10. p–y curves of soil strata at the depth of 0.2 m below mudline.
 D. Mao et al. / Ocean Engineering 100 (2015) 35–45 41

Obviously, the FE model using the lumped parameter model with p–y 4.3. Decay pattern of lateral foundation resistance under cyclic
curve PSI elements has much higher accuracy than that using the loading
equivalent pile method. The simulation results with the lumped
parameter method and the experimental results have fair agreement. With the increase of cycle number, the excess pore water
The calculated natural frequency with p–y curve PSI element is little pressure of the soil around the piles grows up, especially when
bigger than that from the experiment. Some difference may come loading frequency is high it grows fast obviously, as shown in
from the cluster piles effect. If the cluster piles effect were considered Fig. 7. Increasing of the excess pore water pressure is an important
in p–y curve, the deference would be smaller. factor to the soil resistance because it will reduce the soil shear
strength and bearing capacity (Mostafa and El Naggar, 2004).
According to the theory of beam deformation and stress analysis,
the lateral displacement of the model pile and the lateral soil
resistance could be solved from the measured pile strain (Matlock,
Table 7
1970). Fig. 8 is the lateral displacement distribution of the model pile
Stable values of soil resistances at depth 0.2 m below mudline.
during cycle. Inflection point moves toward down with the increasing
T A ¼50 mm A ¼ 100 mm A ¼ 150 mm A ¼200 mm of the displacement of the pile head and the whole length of
(s) deformable pile becomes longer. From the comparisons of the
p (kN/ y p (kN/ y p (kN/ y p (kN/ y
distributions of the soil resistance along the pile before and after
m) (mm) m) (mm) m) (mm) m) (mm)
various cyclic loadings (Fig. 9), the maximum soil resistance appeared
10.0 10.95 2.34 11.94 5.45 11.40 10.86 8.36 20.48 at the site at a distance of 0.2 m below mudline and the attenuation
5.0 10.88 2.31 11.88 5.52 11.41 10.86 8.30 20.42 mainly happened in top clay layer of not more than 1 m deep below
2.0 11.05 2.35 11.82 5.56 11.38 10.83 8.52 20.43 mudline. Fig. 10 demonstrates the changes of p–y curves at the site at a
1.0 11.20 2.31 11.78 5.46 11.56 10.84 9.80 19.16
0.5 11.56 2.30 11.90 5.53 11.86 10.72
distance of 0.2 m below mudline during various cyclic loadings. The
larger the cyclic loading, the greater the soil resistance decay. During

100
acceleration/10-3 g
acceleration/10-3 g

20
5
10
0
0
-10 -50
-20 -100
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
time/s time/s

Fig. 11. Free decay curves. (a) in x direction and (b) in y direction.
Peak acceleration/10-3g

40
30 7.61
20
10
0
0 5 10 15 20 25 30 35 40 45 50
Natural frequency/Hz
Peak acceleration/10-3g
Peak acceleration/10-3g

40 40
30 7.52 30 7.38
20 20
10 10
0 0
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
Natural frequency/Hz Natural frequency/Hz
Peak acceleration/10-3g
Peak acceleration/10-3g

40 40 7.03
30 7.23 30
20 20
10 10
0 0
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
Natural frequency/Hz Natural frequency/Hz

Fig. 12. Impact of load amplitude on fundamental nature frequency of jacket model (bending about x). (a) Before cyclic loading; (b) after cyclic loading with A ¼ 150 mm and
T ¼ 0.5 s; (c) after cyclic loading with A ¼100 mm and T ¼0.5 s; (d) after cyclic loading with A ¼ 50 mm and T ¼0.5 s; (e) after cyclic loading with A ¼200 mm and T ¼ 0.5 s.
42 D. Mao et al. / Ocean Engineering 100 (2015) 35–45

the first 10 cycles the soil resistance decayed very seriously and greater than that under low frequency loading. Reese (1974) and El
reached stable within 100 cycles, which agrees with Matlock field Naggar and Novak (1996) studied damping effect on the soil
experiment in Sabine and Lake Austin (Matlock, 1970). However, when resistance in detail. Table 7 gives the stable values of the soil
the pile displacement reached 0.02 m (0.2D) the phenomenon that the lateral resistances at the depth 0.2 m below mudline under various
soil resistance will decay continuously that Matlock said did not loading cycles. The experimental result shows that the soil
occurred in present experiment, but ran up to stable after 25 cycles. resistance attenuation happens under cyclic loading. The degrada-
Under small amplitude cyclic loading, variation of the soil tion is mainly controlled by the pile motion amplitude. The larger
resistance caused by deferent loading frequencies is small. When the pile motion amplitude, the greater the soil resistance decay.
cyclic loading amplitude is great the high loading frequency will
result in increase of the soil resistance. Loading with large
amplitude and high frequency (2 Hz), though the excess pore 4.4. Pulse exciting and frequency response
water pressure increases obviously and the soil shear strength
depresses, the damping will be enlarged at same time because of Several free decay curves in x and y directions at pipe joints of
the pile velocity increase, which finally makes the soil resistance the upper structure that were low-pass filtered are shown in
Peak acceleration/10 g
-3

30
20 9.51
15
10
5
0
0 5 10 15 20 25 30 35 40 45 50
Natural frequency/Hz
Peak acceleration/10-3g
Peak acceleration/10-3g

30
30
20 9.27
20 9.45 15
15
10 10

5 5
0 0
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
Natural frequency/Hz Natural frequency/Hz
Peak acceleration/10-3g

Peak acceleration/10-3g

30 30 8.87
20 9.10 20
15 15
10 10
5 5
0 0
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
Natural frequency/Hz Natural frequency/Hz
Fig. 13. Impact of load amplitude on fundamental nature frequency of jacket model (Torsion about z axis). (a) Before cyclic loading; (b) after cyclic loading with A ¼ 200 mm
and T ¼ 0.5 s; (c) after cyclic loading with A ¼ 150 mm and T ¼0.5 s; (d) after cyclic loading with A ¼ 100 mm and T¼ 0.5 s; (e) after cyclic loading with A ¼50 mm and T ¼ 0.5 s.

7.7 9.8
T=10.0 s
7.6 7.6 T=10.0 T=5.0 s
T=5.0 s 9.6 9.6 T=2.0
Natural frequency/Hz

Natural frequency/Hz

7.5 1 T=2.0 T=1.0 s

T=1.0 s 1 T =0.5s
7.4 T =0.5s 9.4
7.3
7.2 9.2

7.1
9.0
7.0
6.9 8.8
0 50 100 150 200 0 50 100 150 200
Amplitude/mm Amplitude/mm

Fig. 14. Changes of #1 and #3 nature frequencies after cyclic loading. (a) Bending about x and (b) torsion about z.
 D. Mao et al. / Ocean Engineering 100 (2015) 35–45 43

Fig. 11. These curves were acquired by the pulse excitations with a 50 mm to 200 mm step by step. The 1st and 3rd natural frequencies
force hammer. declined with cyclic loading amplitude increased.
The following equation is used to estimate frequency response Obtained from the experiment, the 1st and 3rd natural fre-
function: quencies of the prototype jacket before and after various cyclic
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi loadings are shown in Fig. 14. From which we see that the
H ν ¼ GðY; YÞ=GðX; XÞ ð3Þ
maximum decrease for 1st and 3rd natural frequencies were only
where, GðY; YÞ is output signal spectrum, GðX; XÞ is input signal 1.3% and 1.9%, respectively, under the cyclic loadings with same
spectrum. amplitude and different frequencies, whereas the correspondingly
The frequency responses of the scale jacket to various pulse reduced values under the cyclic loadings with different amplitudes
excitations before and after cyclic loading with different amplitudes and same frequency were at least 2.3% and 3.2%, respectively. After
and periods are shown in Figs. 12 and 13. The 1st and 3rd natural the last cyclic loading with amplitude A ¼200 mm, period T ¼0.5 s,
frequencies declined successively from 7.61 Hz and 9.51 Hz to 7.03 Hz and cycle number N ¼200 acted on, the 1st and 3rd natural
and 8.87 Hz, respectively, when loading amplitude increased from frequencies of the prototype jacket have fell 7.6% and 7.7% than

Fig. 15. Displacement of pile at mudline.

44 D. Mao et al. / Ocean Engineering 100 (2015) 35–45

Calculation Experiment

A=200mm T=0.5s
displacement at mudline/mm

time/s

Calculation Experiment

Fig. 15. (continued)

5 loading direction at mudline. Fig. 15 shows the comparison of the

y = -1.998t +4.225 numerical simulating and experimental results of the pile displace-
ment at mudline, from which the excellent fitness is verified. During
Peak acceleration/10-3g

4
the first 10 cycles, the tested soil resistance appeared obvious decline.
But the numerical simulating results have no such a distinct phen-
3
omenon because the PSI element in numerical simulation directly
adopts the stable foundation stiffness after degradation. Though
2 there is a little difference between the tested soil resistance and
the numerical result obtained with lumped parameter model at
ω =7.61Hz, ξ=4.1% initial several cycles, the maximum stable experimental and numer-
1
ical soil resistances are almost completely identical.
0
0.0 0.2 0.4 0.6 0.8
Time/s
4.6. Free decay and damping ratio
Fig. 16. Fitting of the amplitude.

Using a free decay curve one is able to estimate the damping ratio
those before cyclic loading affected on. This means that the ξ, and the fundamental natural frequency ω, The peak values An and
influence of the amplitude of marine environment loads on the the frequency of vibration for the different cycles meet the well-
natural frequency of the offshore structure will be much greater known linear free damping decaying Eqs. (4) and (5) (Elshafey et al.,
than that the frequency of marine environment loads. In other 2009). Thus the damping ratio ξ could be estimated by fitting the
words, since the natural frequency of the offshore structure falls raw data curve to the equations. One fitted free decay curve obtained
with the foundation stiffness decrease, the impact of cyclic loading from the free decay test in direction x is given in Fig. 16. The damping
amplitude on the foundation degradation is larger than that of ratio under cyclic loadings ranges from 3.6% to 4.4%. The foundation
cyclic loading frequency. Therefore, for the marine structures decay has little effect on the structure damping ratio.
under longtime large cyclic loads, the effect of foundation degra- 2 3
dation should be pay enough attention. qffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffi
6 ξ 2 7
y ¼ A0 e  ξωt 4 cos ðω 1  ξ Þt þ qffiffiffiffiffiffiffiffiffiffiffiffiffi sin ðω 1  ξ Þt 5
2
ð4Þ
1ξ
2
4.5. Pile displacement in loading direction at mudline

For the pile–soil–upper structure system, the upper structure is

2πξ
made of steel, therefore its mechanical behavior is linear. The motion ln An þ 1 ¼ ln An  qffiffiffiffiffiffiffiffiffiffiffiffiffi ð5Þ
1ξ
2
of the upper-structure is mainly affected by the pile displacement at
mudline. During cyclic loading, the loads on the piles from super-
structure are undertaken by the soil around the piles. The soil where A0 is the initial amplitude, ω is the frequency, and ξ is the
bearing capacity could be represented by the pile displacement in damping ratio, and An is the amplitude after n cycles.
 D. Mao et al. / Ocean Engineering 100 (2015) 35–45 45

4.7. Revision of the pile–soil interaction model Acknowledgment

Since 1975 the p–y curve for the internal domain of the soil of Project is supported by the National High Technology Research
p ¼ 0:5P u ðy=yC Þ1=3 had been adapted in API standard API-RP2A, and Development Program of China (863 Program, Grant no.
many researchers have revised it. In consideration of the founda- 2008AA092701-3), the National Natural Science Foundation of
tion degradation, this paper revised API p–y curve with a reduction China (Grant no. 51379214).
factor δ, which is a function of the cycle number N, the soil elastic
modulus Es, the soil undrained shear strength Cu, the diameter of
the pile D and the lateral pile displacement y, as shown in Eq. (6). References
Taking the tested data under cyclic loadings into the pile–soil
interaction model, i.e. the Eq. (6), one can obtain the relative Adrezin, R., Bar-Avi, P., Benaroya, H., 1996. Dynamic response of compliant offshore
structures – review. J. Aerosp. Eng. 9 (4), 114–131.
coefficients. Since the sand displacement yielded in present American Petroleum Institute, 2000. Recommended practice for planning, design-
experiment is small, only the clay p–y curve revision is carried ing and constructing fixed offshore platforms—Working Stress Design, Twenty-
out with the clay decay pattern under cyclic loading. Fitting the first Edition, December 2000 Errata And Supplement 1, December 2002RP2A-
WSD [S], Dallas, [s. n.].
tested data into p–y curve yields Asgarian, B., Lesani, M., 2009. Pile–soil–structure interaction in pushover analysis of
 1=3  1=3 jacket offshore platforms using fiber elements. J. Constr. Steel Res. 6, 209–218.
y y 0:1E s y
p ¼ 0:5δP u ¼ 0:5½1  0:3ð1  N  0:02370:2D Cu ÞP u ð6Þ CCS, 1992. Rules for the Classification and Construction of Fixed Offshore Platforms.
yc yc China Communications Press, Beijing, In Chinese.
Dongfeng, Mao, Minghui, Zhang, Yang, Yu, Jun, Zhao, 2015. Analysis of spudcan–
where, Pu is the ultimate bearing capacity of clay, yc ¼ 2.5εcD, and footprint interaction in a single soil with nonlinear FEM. Pet. Sci. 12 (1),
εc is the pile strain when the half maximum stress appeared for 148–156.
Elshafey, Ahmed A., Haddara, Mahmoud R., Marzouk, H., 2009. Dynamic response
undisturbed soil in undrained test.
of offshore jacket structures under random loads. Mar. Struct. 22, 504–521.
El Naggar, M.H., Novak, M., 1996. Nonlinear analysis for dynamic lateral pile
response. Soil Dyn. Earthq. Eng. 15 (4), 233–244.
5. Conclusions El Naggar, M.H., Bentley, K.J., 2000. Dynamic analysis for laterally loaded piles and
dynamic p–y curves. Can. Geotech. J. 37 (6), 1166–1183.
Gudmestad, O.T., Moe, G., 1996. Hydrodynamic coefficients for calculation of
Based on investigation about the lateral vibration characteristics of
hydrodynamic loads on offshore truss structures. Mar. Struct. 9 (8), 745–758.
the offshore jacket involving in the foundation degradation using scale Horr, A.M., Safi, M., 2003. Full dynamic analysis of offshore platform structures
model experiment and FE analysis with the model designed in present using exact Timoshenko pipe element. J. Offshore Mech. Arct. Eng. 125 (3),
168–175.
paper, some valuable conclusions are obtained.
Karunakaran D., Baerheim M., Leira B.J., 1997. Measured and simulated dynamic
response of a jacket platform. In: Proceedings of the Sixteenth Offshore
(1) The water depth has a little effect on the natural frequency of Mechanics and Arctic Engineering Conference II, pp. 157–164.
the offshore jacket. Li, Zhi-gang, Yuan, Zhi-lin, Duan, Meng-lan, et al., 2012. Development of pile–soil
interaction test system for jacket platform and its application. Rock Soil Mech.
(2) In saturated soil, the reciprocating motion of piles induces the 33 (12), 3833–3840, In Chinese.
excess pore water pressure of the foundation to go up and the Ling, Xian-chang, Wang, Chen, Wang, Cheng, 2004. Scale modeling method of
lateral soil resistance to decay while the weakened constraint of shaking table test of dynamic interaction of pile–soil–bridge structure ground
of soil liquefaction. Chin. J. Rock Mech. Eng. 23 (3), 450–456.
the foundation makes the decrease of the natural frequency of the Matlock H., 1970. Correlations for design of laterally loaded piles in soft clay. In:
offshore structure. The more the soil resistance decaying, the Proceedings of the 2nd Offshore Technology Conference, Houston, [s. n.].
greater the natural frequency of offshore jacket attenuates. The Mostafa, Y.E., El Naggar, M.H., 2004. Response of fixed offshore platforms to wave
and current loading including soil–structure interaction. Soil Dyn. Earthq. Eng.
decrease of the natural frequency is mainly affected by the cyclic
24, 357–368.
loading amplitude but the cycle frequency. However, the relation Nogami, T., Novak, M., 1980. Coefficients of soil reaction to pile vibration. J. Geotech.
between the natural frequency attenuation and the soil resistance Eng. 106 (5), 565–570.
decaying need to be studied further. Nogami, T., Konagai, K., 1988. Time-domain flexural response of dynamically loaded
single piles. J. Eng. Mech. 114 (9), 1512–1525.
(3) The comparisons among experimental results and two groups Novak, M., Nogami, T., Aboul-Ella, F., 1978. Dynamic soil reaction for plane strain
of the finite element simulating results were conducted. The case. J. Eng. Mech. 104 (4), 953–955.
relative differences between the simulating results of the Ou, J., Long, X., Li, Q.S., Xiao, Y.Q., 2007. Vibration control of steel jacket offshore
structures with damping isolation systems. Eng. Struct. 29, 1525–1538.
lumped parameter method with p–y curve PSI element and Reese L.C., Cox W.R., Koop F.D., 1974. Analysis of laterally loaded pile in sand. In:
the experimental results locate from 5.2% to 7.8%. The fair Proceedings of the 6th Offshore Technology Conference, Houston, [s. n.].
agreement was obtained. But the relative differences between Shang, Shou-ping, He, Zhi-wen, Wang, Hai-dong, et al., 2008. Experimental
investigation on the effect of the relative stiffness ratio between
the simulation results of the equivalent pile method and the
superstructure and ground soil on the fundamental frequency of soil–
experimental results are more than 20%. The maximum structure system. J. Earthq. Eng. Eng. Vib. (5), 94–101.
difference reaches 29%. Therefore, it is suggested that it would Sunder, S.S., Connor, J.J., 1981. Sensitivity analyses for steel jacket offshore plat-
be much better to use the lumped parameter model with p–y forms. Appl. Ocean Res. 3 (1), 13–26.
Wang, Jian-hua, Feng, Shi-lun, 2005. Research on lateral resistance of pile
curve PSI element to analyze dynamic behaviors of an offshore foundation in liquefaction strata. Rock Soil Mech. 26 (10), 1597–1601, In
structure in real engineering. Chinese.
(4) In presupposition of proportional damping, the investigation of the Wisch D.J., 1998. Fixed steel offshore structure design – past, present & future. In:
Proceedings of the Offshore Technology Conference. Houston, Texas, 4–7 May
peak value of free decay vibration shows that the impact of the 1998, pp. 317–323.
foundation degradation on the system damping is very small.
(5) A revised pile–soil interaction model, i.e. a revised p–y curve,
taking the cycle number and lateral pile displacement into
consideration is suggested.