Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 45 No.

4 December 2014 ISSN 0046-5828

Predicting Spud Can Extraction Resistance in Soft Clay

Omid Kohan1, Christophe Gaudin2, Mark J. Cassidy3, and Britta Bienen4
1
Centre for Offshore Foundation Systems and ARC Centre of Excellence for Geotechnical Science and Engineering, The
University of Western Australia, Perth, Australia
2
Centre for Offshore Foundation Systems and ARC Centre of Excellence for Geotechnical Science and Engineering, The
University of Western Australia, Perth, Australia
3
Centre for Offshore Foundation Systems and ARC Centre of Excellence for Geotechnical Science and Engineering, The
University of Western Australia, Perth, Australia
4
Centre for Offshore Foundation Systems and ARC Centre of Excellence for Geotechnical Science and Engineering, The
University of Western Australia, Perth, Australia
1
E-mail: [email protected]
2
E-mail: [email protected]
3
E-mail: [email protected]
4
E-mail: [email protected]

ABSTRACT: Jack-ups are mobile offshore structures that are frequently relocated to new operation sites. To be relocated, the jack-up
footings, known as spudcans need to be extracted from the seabed, using essentially the buoyancy of the hull as extraction force. This
operation may be time consuming or even jeopardised if the spudcan extraction resistance is higher than the available extraction force. The
maximum extraction (or breakout) resistance consists of suction at the spudcan base, weight of the soil above the spudcan, and soil shear
resistance above the spudcan, with the contribution of the suction at the spudcan invert being the dominant component of the breakout
resistance. This paper reviews an existing prediction method used to estimate spudcan extraction resistance and proposes an update of some
of the input parameters based on insights obtained from a large database of experimental model data on two types of clays and for spudcan
embedment up to three diameters.

KEYWORDS: Spudcan, Extraction resistance, Centrifuge modelling, Soft clay, Prediction method

1. INTRODUCTION Two methods have been developed to estimate the maximum

spudcan extraction resistance. They are detailed in Purwana et al.
Self-elevating mobile jack-up units are the most common facilities (2009) and Osborne et al. (2011), respectively. The method detailed
used for offshore drilling operations in shallow waters, up to in Purwana et al. (2009) is based on measurements of total and pore
approximately 150 m depth (Figure 1). Once operation is completed, pressure at various locations on a model spudcan in centrifuge
the jack-up is relocated to a new operation site, necessitating the experiments for embedment up to 1.5 spudcan diameters, as well as
jack-up footings, known as spudcans, to be extracted from the information regarding soil failure mechanism from Particle Image
seabed. Difficulties in extraction can arise if the spudcans are deeply Velocimetry analysis. The method contained in Osborne et al.
embedded in very soft clays. The development of high suction (2011) is a modified version of this.
forces at the spudcan invert (Purwana et al., 2005; Gaudin et al., The objective of this paper is (i) to check the validity of the
2011) may augment the extraction resistance beyond the extraction method established by Purwana et al. (2009) (called here after the
force generated by the hull buoyancy, resulting in unexpected delays reference method) for spudcan embedment up to 3 diameters and (ii)
and additional costs. presents an update of some of the input parameters, based on
As a part of assessing the jack-up removal process prior to going insights obtained from an experimental model database of 24
on a new location, an estimation of spudcan extraction resistance is centrifuge tests featuring spudcan extraction from normally
therefore necessary for the jack-up operators to anticipate potential consolidated clay.
extraction issues and develop mitigation measures to facilitate
spudcan extraction, such as water jetting for instance (Bienen et al.,
2. DATABASE
2009; Gaudin et al., 2011).
The experimental database was gathered from data reported by
Purwana et al. (2005), Purwana et al. (2009), Gaudin et al. (2011),
Kohan et al. (2013a), Kohan et al. (2013b), and Kohan et al. (2014).
A total of 24 centrifuge test results were extracted and they are
summarised in Table 1 in prototype scale. Scale factors for
geometry, load, pressure, and the diffusion process can be found at
Garnier et al. (2007) who made an inventory of the scaling laws and
Over similitude questions relating to centrifuge modelling. Tests were
160m conducted at 100 and 200 g, modelling spudcans of 6, 8, 12.5 and
17.1 m in diameter (30, 40, 85.56 and 125 mm in model scale).
4m For all tests, the test procedure consisted of three stages. In the first
stage, spudcan penetration was performed in-flight in displacement
Over or load control, under undrained conditions. The spudcan
20m All dimensions 20m installation depths varied from 1 to 3 times the spudcan diameter. In
approximate
the second stage, the jack-up operation period was simulated by
maintaining a constant vertical load between 50% and 90% of the
Figure 1 Typical jack-up and spudcan (modified after Reardon,
maximum installation load for up to five years in prototype scale,
1986)
achieving varying degrees of consolidation in the soil around the

53
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 45 No.4 December 2014 ISSN 0046-5828

spudcan. It is noteworthy that the effect of operation load is less This has been identified by both PIV analysis (Purwana et al.,
significant than that of operation duration (Purwana et al., 2005). 2006a) and numerical analysis (Zhou et al., 2009) of spudcan
Finally, in the third stage, spudcan extraction was performed in extraction in normally consolidated clay. Kohan et al. (2013b)
displacement control at a rate, v, resulting in a normalised velocity demonstrated that this mechanism is also relevant for initial
V=vD/cv greater than 30, where cv is the virgin consolidation embedment ratio up to 3 times the spudcan diameter.
coefficient and D is spudcan diameter. This ensured that spudcan The components involved in the spudcan extraction resistance
extraction was also performed under undrained conditions (Finnie are influenced by the duration of the jack-up operation, i.e. by the
and Randolph, 1994). The maximum extraction loads are reported in degree of dissipation of excess pore pressures generated during
Table 1. installation, in the soil surrounding the spudcan. This results in the
Centrifuge studies on spudcan extraction employed for shear strength of the soil surrounding the spudcan increasing with
assessment of the spudcan extraction resistance were performed in operation time, and consequently, an increase in extraction
two different soils: UWA kaolin clay and Malaysian kaolin clay resistance, as already demonstrated by Purwana et al. (2005).
with a coefficient of consolidation cv of approximately 2.8 to 4.8 It is noted that this mechanism may not apply for spudcans that
m2/year for UWA Kaolin clay and 40 m2/year for Malaysian kaolin have not seen any dissipation of excess pore pressures at immediate
clay at a stress level consistent with the spudcan embedment. Soil extraction. In this case, a reverse flow mechanism is more likely to
characteristics including soil shear strength, soil unit weight, and develop.
soil effective stress at the spudcan installation depth for each
centrifuge test are also provided in Table 1. 4. EVALUATION OF THE REFERENCE METHOD

3. EXTRACTION FAILURE MECHANISM The method proposed by Purwana et al. (2009) (reference method)
is based on the aforementioned breakout failure mechanism,
The spudcan extraction failure mechanism was described in detail identified using Particle Image Velocimetry (PIV) analysis for
by Purwana et al. (2009) and Gaudin et al. (2011) for embedment undrained extraction of a 12.5 m in diameter spudcan (prototype
ratios up to 1.5 times the spudcan diameter. The mechanism at peak scale) from a depth of approximately 1.5 spudcan diameters in
extraction resistance is a combination of an uplift mechanism of the Malaysian kaolin clay (Purwana, 2006b). The vertical uplift force
soil at the top of the spudcan, and a reverse end bearing at the equilibrium condition assumed by Purwana et al. (2009) is
spudcan invert associated with the development of negative excess illustrated in Figure 2.
pore pressure, namely suction (Figure 2). The main soil resistance is The method has been presented in details in Purwana et al.
comprised of the weight of the soil above the spudcan, the resistance (2009). It computes the uplift resistance as the sum of a resistance at
along a shear plane generated above the spudcan, and the suction the base Qbase (which accounts for overburden stresses), at the top
pressure at the spudcan base. Qtop and the submerged weight of the spudcan Weff. Table 2 (see
also Figure 3 and Figure 4) details the calculation of the first two
components and summarises the parameters used in the method.
When determining the net extraction resistance, Weff is considered
as zero.
To evaluate the performance of the method, the peak extraction
resistance was calculated for each case, based on the input
parameters reported in Table 3. Additional assumptions were made
when data were missing, as explained below:

Figure 3 Variables defined in Table 2

i) To compute the top soil resistance, the height of the soil flowing
back onto the top of the spudcan, which is a function of the depth of
cavity formed during deep installation, needs to be assessed For
Figure 2 Observed spudcan breakout failure mechanism and cases where the cavity depth Hc was not reported, the solution
diagram of breakout force components (after Purwana et al., 2009) developed by Hossain et al. (2006) was used as expressed in Table
2.
54
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 45 No.4 December 2014 ISSN 0046-5828

Table 1 Database

Test Reference Ratio of the Spudcan Spudcan Operation Operation Breakout Soil Soil shear Soil
name centrifugal diameter depth time load level load unit strength at effect
acceleration ratio of the weight installation depth ive
to the earth maximum stress
gravity installation
load
D H/D t Vop/Vp Qc ' su 'v
(m) (-) (day) (-) (MN) (kN/m3) (kPa) (kPa)
Kohan et al.
3.0D2.0Y 200 6.00 3.02 730 85% -6.14 6.20 1.10 H 112.41
(2013b)
Kohan et al.
2.5D2.0Y 200 6.00 2.50 730 85% -5.62 6.20 1.10 H 93.06
(2013b)
Kohan et al.
2.0D2.0Y 200 6.00 1.99 730 85% -4.13 6.05 1.10 H 72.06
(2013b)
Kohan et al.
1.5D2.0Y 200 6.00 1.48 730 85% -3.29 6.05 1.10 H 53.54
(2013b)
Kohan et al.
1.5D3.0Y 200 6.00 1.48 1095 85% -3.43 6.05 1.10 H 53.66
(2013b)
Kohan et al.
1.5D1.0Y 200 6.00 1.47 365 85% -2.82 6.05 1.10 H 53.48
(2013b)
Kohan et al.
1.5D0.5Y 200 6.00 1.47 183 85% -2.42 6.05 1.10 H 53.48
(2013b)
Kohan et al.
1.5D0.0Y 200 6.00 1.47 5 0% -1.58 6.05 1.10 H 53.48
(2013b)
Kohan et al.
Nojet2 200 8.00 3.02 730 85% -13.91 7.50 1.04 H 180.79
(2014)
Kohan et al.
Nojet1 200 8.00 2.50 730 85% -14.24 7.50 1.08 H 182.01
(2013a)
Gaudin et al.
S1UEnJ 200 17.11 1.46 1664 90% -80.52 6.00 1.17 H 150.00
(2011)
Gaudin et al.
S2UEnJ 200 17.11 1.05 1664 90% -46.97 6.00 1.17 H 108.00
(2011)
Purwana et al.
GS1 100 12.50 1.45 <1 0% -17.57 6.50 1.56 H 117.65
(2005)
Purwana et al.
GS2 100 12.50 1.51 53 75% -19.60 6.50 1.56 H 122.85
(2005)
Purwana et al.
GS3 100 12.50 1.51 126 75% -24.62 6.50 1.56 H 122.85
(2005)
Purwana et al.
GS4 100 12.50 1.47 244 75% -27.69 6.50 1.56 H 119.60
(2005)
Purwana et al.
GS5 100 12.50 1.52 423 75% -31.26 6.50 1.56 H 123.50
(2005)
Purwana et al.
GS6 100 12.50 1.50 843 75% -36.19 6.50 1.56 H 122.20
(2005)
Purwana et al.
D-01 100 12.50 1.21 400 50% -23.57 6.50 1 + 1.60 H 98.54
(2009)
Purwana et al.
D-02 100 12.50 1.51 400 50% -31.89 6.50 1 + 1.60 H 122.85
(2009)
Purwana et al.
D-03 100 12.50 1.77 400 50% -37.93 6.50 1 + 1.60 H 143.59
(2009)
Purwana et al.
C-03 100 12.50 1.71 <1 0% -20.12 6.50 1 + 1.60 H 138.65
(2009)
Purwana et al.
C-02 100 12.50 1.51 <1 0% -19.05 6.50 1 + 1.60 H 122.66
(2009)
Purwana et al.
C-01 100 12.50 1.45 <1 0% -17.39 6.50 1 + 1.60 H 117.59
(2009)

55
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 45 No.4 December 2014 ISSN 0046-5828

Table 2 Parameters of reference method and method of this study

Parameter Description Reference method Updated Method Comments

D Spudcan diameter (m)

H Spudcan installation depth (m)
Hc Cavity depth (m) Measured during centrifuge test H c  suHc 
0.55
1  s  Provides
    uHc  universal method
D   '.D  4   '.D  of all clay
conditions
Hs Spudcan side wall (m)
Ht Height of backfill above spudcan See Figure 3
top surface (m)
'top Unit weight of soil at top (kN/m3) 'top =0.92 ' 'top =' Simplification
required without
reliable method
to estimate the
change in ' 
' Unit weight of soil (kN/m3) Soil property
su,top Average shear strength of backfill
soil above spudcan after
installation (kPa)
su,base Shear strength at the spudcan
base level after installation (kPa)
fg,top Change in shear strength of soil 0.67 for immediate extraction Figure 8 Provide
above spudcan top due to soil 0.87 for 400 days operation estimation for the
disturbance and any soil full range of
reconsolidation after spudcan operational
installation (-) periods
fg,base Gain in shear strength of soil 1.00 for immediate extraction Figure 9 Provide
below spudcan base due to any 1.70 for 400 days operation estimation for the
soil reconsolidation after spudcan full range of
installation (-) operational
periods
S Shape factor See Figure 4
Nc,top Breakout factor for top soil Ht   H   top
'
.H t
resistance1 (-) if  1 S * 2.56 ln 2 t   
D   D  ( s u ,top ave . f g , top
)
H   H t   top .H t
'

if t  1 3.5489 *    1
D   D  ( s u ,top ) ave . f g ,top

Nc,base Breakout factor for base soil   H 

resistance (-) N c , base  4 . 1  0 . 2     5 .2
  D 
Sb Adjustment factor for overburden  H  1.00 for all cases; Simplification
 0.00 for  0.35 
stress at spudcan base level D therefore, is not part of required without
 H H  the updated method reliable method
Sb  0.87  0.305 for 0.35   1.5
 D D  to estimate the
 H change of Sb
1 .0 for 1.5   2 .0 
 D 
Quplift Total uplift resistance Q uplift  ( Q top  Q base  W eff )
Qtop Top soil resistance Qtop  0.25D 2 ( N c,top .su ,top ) * f g ,top
2
Qbase Base soil resistance Qbase  0.25D 2 ( N c,base.su,base. f g ,base   '.H. Qbase  0.25D ( f sr .N c ,base .su ,base . f g ,base . f ol   ' H
2

To consider effect of operation load and

strength ratio
Weff Submerged weight of spudcan Weff is ignored for net uplift resistance
fol Factor of operation ratio Vop 0.5 ≤ Vop/Vp ≤ 1
f ol  1  0.2(2.  1)
Vp

fsr Factor of strength ratio  v' Best fit to

f sr  1  0.4(  1) conditions of
4su database

1. It should be noted that in calculation of breakout factor for top soil resistance, parameter fg,top was not mentioned in the original
formulae presented by Purwana et al. (2009) and Purwana et al. (2010). However, the first author was informed by the personal
correspondence that it is included in the overburden pressure term for determination of top soil resistance breakout factor (Purwana, 2010).
2. It is noted that for immediate or short consolidation periods fg,base will be 1 or close to 1 and Qbase can predict a negative value. This is
inappropriate and reflects that the mechanism for this case is not realistic.

56
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 45 No.4 December 2014 ISSN 0046-5828

period. This resulted in values of 0.67 and 0.87 recommended by

Purwana et al. (2009) for 0 and 400 days of operation time,
respectively. Similarly to the calculation of fg,base, linear
interpolation was conducted to assess fg,top for intermediate
consolidation times.
From Figure 6, showing pore pressure responses at the end of
the installation and operation time with respect to the hydrostatic
pressure for test GS5, a degree of consolidation of 41% was deduced
for 400 days of operation time. Calculated values of fg,top are
presented in Table 3.

Figure 4 Shape factor (after Merifield et al., 2003)

0.55
H c  suHc  1 s  (1)
    uHc 
D   '.D  4   '.D 

where Hc = Cavity depth (m); D = Spudcan diameter (m); ’  =

Effective unit weight of soil (kN/m3); suHc = Shear strength at the
cavity depth (kPa).
ii) The gain in soil shear strength underneath the spudcan during
operation time is characterised by the parameter fg,base, which was
evaluated as 1.00 and 1.70 by Purwana et al. (2009) (from
numerical analysis) for immediate extraction and extraction after
400 days of operation, respectively. No values were reported for Figure 6 Pore pressure responses at spudcan top (after Purwana et
intermediate operational times. To evaluate the performance of the al., 2005)
reference method for the entire database, presented in Table 1, fg,base
was calculated for intermediate operation times, by linear Predictions from the reference method are compared with the
interpolation between the degrees of consolidation achieved for 0 measured uplift resistances in Figure 7. Two observations are made:
and 400 days. At 0 days operation time, it is logical to assume that 1. The method predicts reasonably the peak extraction resistance
the degree of consolidation is equal to 0. For 400 days and for in Malaysian clay (which is expected as the data underpin the
Malaysian clay, the excess pore pressure dissipation during development of this method), with a mean percentage error of about
operation time was not reported in Purwana et al. (2009). 9%. The performance is reduced for UWA kaolin clay, with a mean
Accordingly, data reported in Purwana et al. (2005) for test GS5 percentage error of about 57%. This potentially indicates that the
(423 days operation time) was used (Figure 5), leading to a degree performance of the reference method may be affected by the nature
of consolidation of 78% at 400 days. For tests in UWA kaolin clay, of the clay and that a better understanding of the various parameters
the degrees of consolidation were extracted from pore pressure associated with the soil characteristics is required.
measurements. Results of the linear interpolation of fg,base are listed 2. The performance of the reference method is consistent for
in Table 3. both clays for spudcan embedment up to 3 diameters, extending the
validity of the method from embedment of 1.5 to 3 spudcan
diameters. This is consistent with findings from Kohan et al. (2014),
which demonstrated that the failure mechanism during extraction
was identical between embedment of 1.5 and 3 spudcan diameters.

Figure 5 Dissipation of excess pore pressure at spudcan base during

operation period (after Purwana et al., 2005)

iii) The change in soil shear strength at the top of the spudcan
due to installation is characterised by the parameter fg,top. Purwana et
al. (2009) performed a series of T-bar tests in Malaysian kaolin clay
to measure the shear strength of the remoulded soil at the spudcan
top during the operation period. The shear strength was measured to
reduce to 67% of the undisturbed shear strength immediately after
spudcan installation, but increased by 30% (or 87% of the Figure 7 Predicted uplift force based on the reference method
undisturbed soil shear strength) after 400 days reconsolidation proposed by Purwana et al. (2009)
57
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 45 No.4 December 2014 ISSN 0046-5828

Table 3 Performance of the reference method

Test name Cavity Soil Soil Shear Shear Bearing Bearing Top Base Predicted Measured Error
depth shear shear strength strength factor at factor at resistance resistance breakout breakout
strength strength gain at gain at top base
at top at base top base
- Dc (m) su,top su,base fg,top fg,base Nc,top Nc,base (-) Qtop Qbase Qbreakout Qc -
(kN/m2) (kN/m2) (-) (-) (-) (MN) (MN) (MN) (MN) (%)
3.0D2.0Y 0.43 9.16 19.94 0.92 1.45 12.56 5.20 -3.25 -1.07 -4.32 -6.14 -42.05
2.5D2.0Y 0.43 8.03 16.51 0.97 1.54 12.56 5.20 -2.85 -1.12 -3.97 -5.62 -41.56
2.0D2.0Y 0.43 6.09 13.10 0.93 1.47 12.56 5.20 -2.16 -0.79 -2.95 -4.13 -39.91
1.5D2.0Y 0.43 4.25 9.74 0.87 1.36 12.56 5.18 -1.51 -0.46 -1.97 -3.29 -67.33
1.5D3.0Y 0.43 4.50 9.76 0.92 1.45 12.56 5.18 -1.60 -0.58 -2.18 -3.43 -57.27
1.5D1.0Y 0.43 3.97 9.72 0.81 1.26 12.56 5.18 -1.41 -0.31 -1.72 -2.82 -63.78
1.5D0.5Y 0.43 3.50 9.72 0.72 1.08 12.56 5.18 -1.24 -0.07 -1.31 -2.42 -85.23
1.5D0.0Y 0.43 3.27 9.72 0.67 1.00 12.56 5.18 -1.16 0.05 -1.11 -1.58 -42.56
Nojet2 0.46 11.04 25.14 0.89 1.39 12.56 5.20 -6.97 -0.06 -7.04 -13.91 -97.78
Nojet1 0.46 11.46 26.21 1.01 1.62 12.56 5.20 -7.23 -0.34 -7.58 -14.24 -87.90
S1UEnJ 2.09 13.29 29.25 0.87 1.36 12.56 5.17 -38.38 -13.80 -52.18 -80.52 -54.32
S2UEnJ 2.21 9.80 21.06 0.87 1.36 11.53 4.84 -26.92 -16.63 -43.55 -46.97 -7.86
GS1 1.92 9.42 28.24 0.67 1.00 12.56 5.16 -14.51 -4.09 -18.60 -17.57 5.56
GS2 1.92 10.55 29.48 0.72 1.09 12.56 5.20 -16.27 -5.41 -21.68 -19.60 9.59
GS3 1.92 11.63 29.48 0.79 1.22 12.19 5.20 -17.41 -7.92 -25.32 -24.62 2.78
GS4 1.92 12.03 28.70 0.84 1.31 11.52 5.18 -17.00 -9.59 -26.59 -27.69 -4.12
GS5 1.92 12.84 29.64 0.87 1.36 11.55 5.20 -18.19 -10.61 -28.80 -31.26 -8.53
GS6 1.92 12.79 29.33 0.88 1.37 11.41 5.20 -17.91 -10.70 -28.61 -36.19 -26.48
D-01 2.50 11.77 25.26 0.87 1.36 8.86 4.97 -12.79 -11.92 -24.71 -23.57 4.61
D-02 2.50 14.37 31.24 0.87 1.36 10.27 5.20 -18.12 -12.08 -30.20 -31.89 -5.58
D-03 2.50 16.59 36.34 0.87 1.36 11.60 5.20 -23.62 -13.98 -37.60 -37.93 -0.89
C-03 2.50 12.37 35.13 0.67 1.00 12.56 5.20 -19.07 -5.40 -24.47 -20.12 17.78
C-02 2.50 11.05 31.19 0.67 1.00 12.06 5.20 -16.35 -4.85 -21.20 -19.05 10.16
C-01 2.50 10.63 29.94 0.67 1.00 11.65 5.16 -15.20 -5.19 -20.38 -17.39 14.69

5. UPDATING THE INPUT PARAMETERS 5.2 Unit weight of soil above the spudcan, ’top
The reference method is based on a rigorous description of the The unit weight of the soil above the spudcan is slightly lower than
failure mechanism, incorporating the change in strength at the base that of the undisturbed soil due to the heavy remoulding occurring
and top of the spudcan resulting from installation and operation. during penetration. Purwana et al. (2009) assumed the remoulded
They are estimated via two empirical factors fg,base and fg,top, with unit weight of Malaysian clay was about 92% of the virgin soil.
limited insights into the values to adopt for intermediate operational Without indications about the variation of unit weight with the level
times (i.e. between no and full consolidation) and different type of of remoulding or estimated values for other types of clay, it is
clays. The lower performance of the method for kaolin clay suggested for the updated method to use the virgin soil unit weight
indicates that some aspects of the soil characteristics, which are not for all predictions. The impact of such a simplification on the
accounted for in the method, require a closer examination. Potential performance of the method is limited, especially for shallow
candidates include soil sensitivity, undrained bearing capacity penetrations (and limited volume of soil). It is made for ease of
factor, operation load, and consolidation coefficient. calculation.
The database gathered enables additional insights into each
parameters involved in the reference method, although it is noted 5.3 Change in soil shear strength above the spudcan, fg,top
that a range of parameter combinations can be derived to fit
individual test data (non-unique solution) and a holistic view of the During installation, the soil is experiencing heavy remoulding and
fit to the database must be taken. Accordingly, the paper proposes softening, resulting in a reduction of the shear strength of the soil
updated recommendations to estimate the model parameters, notably resting above the spudcan. The reduction in shear strength is a
the values of fg,top and fg,base as a function of the operation time and function of the soil sensitivity St, which may vary between 2 to 2.5
the type of clay used. Two plots are suggested that enable the for UWA kaolin clay and between 2 to 4 for the Malaysian clay.
assessment of gain in shear strength of soil at top and base of Regardless of the soil sensitivity, the remoulded soil regains some of
spudcan depending on the operation time. Additional its shear strength during operation through consolidation.
recommendations relate to the estimation of the cavity depth, the The evolution of the factor fg,top with operation time has been
unit weight of the soil on top of the spudcan, the overburden back calculated from the experimental data presented in Table 1.
adjustment factor, and the introduction of two additional factors to The process results in solving one linear equation with two
account for the effects of the strength ratio and operation load. The unknowns, fg,top and fg,base, requiring additional assumptions on both
updated recommendations for the input parameters are explained in parameters. Accordingly, the following criteria were used to
detail below and are summarised in Table 2. determine soil shear strength at top (and base as explained in next
section) of the spudcan:
5.1 Cavity depth, Hc 1. The lower bound of fg,base is 1, corresponding to the value
immediately after extraction, before any consolidation occurs.
The top soil resistance is a function of the height of the soil flowing 2. The upper bound of fg,top is equal to 1, corresponding to full
back onto the top of the spudcan and of the depth of the cavity strength recovery of the soil above the spudcan after full
formed during installation. The solution developed by Hossain et al. reconsolidation.
(2006) is therefore recommended to estimate the cavity depth as
explained in the previous section.
58
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 45 No.4 December 2014 ISSN 0046-5828

The values of fg,top that were considered a best-fit of the database are data and therefore may be applicable only for Malaysian clay. In the
plotted in Figure 8 against the degree of consolidation U for all tests, present study, to cover all embedment, the two types of clay, and to
excluding tests S1UEnJ, S1UEnJ, GS2 to GS4, and GS6, for which simplify the approach, it is assumed that the overburden stress is
U is unknown. All points fall within in single logarithmic curve fully mobilised at any spudcan embedment depth, resulting in an
demonstrating a loss of strength of 40% immediately after adjustment factor equal to one.
installation and a rapid recovery to a value of about 78-85% of the
initial shear strength after 10% of consolidation. This indicates that
a fairly complex soil hardening process is taking place, which may
involve mechanisms other than consolidation. For prediction
purposes, it is suggested to adopt a value of fg,top of 0.6 for
immediate extraction and a value of 0.78 to 0.85, increasing linearly
between 10 and 100% consolidation. If more detailed knowledge of
an offshore soil is known these limits may potentially be altered.
However, without such knowledge the suggested values are a good
guide that fits the experimental database well.

Figure 9 Gain in shear strength at base of the spudcan during the

operation time

5.6 Effect of the operation load, fol

Purwana et al. (2005) examined the effect of the operational load on
the spudcan extraction. Three tests with the operational load Vop set
at 25%, 50% and 75% of the installation load Vp and with the same
penetration depth and operation period were performed. Comparing
the test results shows that the operation load does not influence the
Figure 8 Change in shear strength at top of the spudcan during top soil resistance, whereas base soil resistance increases by
the operation time approximately 10%, between an operation load ratio of 50% and
75% (Figure 10).
5.4 Gain in soil shear strength underneath the spudcan, fg,base As the reference method was established based on the tests with
The soil below the spudcan consolidates under the load held during an operation load ratio of 50%, a new factor fol (Vop/Vp) with a value
the operational period. This results in a gain in soil shear strength, of 1 at the operation load ratio of 0.5 and upper bound value of 1.2
described by the factor fg,base, which lower bound value is at the operation load ratio of 1 is defined as:
established at 1. Values of fg,base considered to be the best holistic fit
to the centrifuge data (and are consistent with the fg,top values of fol = 1 +  (2 (Vop/Vp) - 1) (2)
Figure 8) are plotted against the degree of consolidation U in Figure
9. This figure covers a wide range of degree of consolidation where Vop = Operation load (MN); Vp = Penetration load (MN);  =
ensuing from different operation periods. It is evident from Figure 9 0.2 (-). The value of  =0.2 has been chosen as it best fits on
that a higher degrees of consolidation results in a larger gain in soil increase of 20% on the base soil resistance, as measured
shear strength beneath the spudcan, with a linear fit reasonably experimentally by Purwana et al. (2005). It should be noted that
representing the data. Values that range from 1 to 1.8 provide a Equation 2 has been fit to the data for 0.25 ≤ Vop/Vp ≤ 1. For Vop/Vp
reasonable fit for the two clays. The linear increase of strength with less than 0.5, a value of 0.9 is recommended.
degree of consolidation is somewhat surprising and potentially
indicates that the gain in strength is not homogenous underneath the
spudcan.

5.5 Overburden pressure adjustment factor, Sb

The weight of the overlying soil imposes an overburden pressure at
the spudcan installation depth. Since the failure mechanism at the
spudcan invert has been identified as a reverse end bearing before
changing to a localised flow around mechanism at the peak
extraction resistance, the overburden pressure is required to be
calculated to determine the net extraction resistance.
Purwana et al. (2009) considered the overburden stress as part of
the base resistance, and assumed that it was partially mobilised from
embedment between 0.35 and 1.5, before being fully mobilised for
Figure 10 Variation of ultimate breakout forces for various ratios of
embedment ranging from 1.5 to 2 (see adjustment factor in Table 2).
operation load to installation load (after Purwana et al., 2005)
This factor was established from back calculation of the centrifuge

59
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 45 No.4 December 2014 ISSN 0046-5828

5.7 Effect of the strength ratio on the breakout factor for base 6. DISCUSSION
soil resistance, fsr
Figure 11 illustrates the performance of the improved method
Figure 8 and Figure 9 demonstrate that a consistent set of through comparison between the predicted and experimental net
parameters could be chosen for representing the consolidation extraction resistance. The method predicts the peak extraction
characteristics and sensitivity of the two soils. However, the resistance equally well in both types of clay (Figure 12), with a
different performance of the two clays still requires differentiation mean difference of about 8% (Table 4).
and a parameter is required to explain the lower performance of the Although the method was used here to simulate test of
reference method for the UWA Kaolin clay. immediate extraction (i.e. no load hold and therefore no excess pore
One component of the increased extraction resistance measured pressure dissipation and consolidation), and was found to provide a
in the UWA tests may be because of the difference in undrained resistance similar to the experiment, it is questioned if the
shear strength profile; with the UWA tests having lower increasing mechanism that the reference method if based on is appropriate for
strength with depth compared to the Malaysian clay tests. Usually this case. It is more likely a localised reverse flow mechanism for
for shallow foundations, considering their behaviour in deep embedments and this mechanism should be the basis of a
compression, this would lead to a lower bearing capacity factor. method to predict immediate extraction.
However, close inspection of the lower bound bearing capacity
factors of Houlsby and Martin (2003) for spudcans in an open cavity
actually shows on increase in the bearing capacity factors with
decreased strength gradient once the spudcan becomes embedded
more than one diameter. As the mechanism of the reference method
has a spudcan bottom contribution as the reverse of the Houlsby and
Martin (2003) solution (i.e. uplift rather than penetration) it can be
assumed that a lower strength gradient can increase the bearing
capacity factor. This is accounted for by introducing a new factor fsr.
An approach is proposed here whereas the effect of soil strength
defined as the soil shear strength normalised by effective stress is
considered. As the reference method was developed based on the
results of the centrifuge tests in Malaysian clay, an additional factor
fsr (’v / su) is defined as a function of the ratio of the effective stress
normalised by soil shear strength for any soft soils to that of
Malaysian kaolin clay. The effective stress normalised by soil shear
strength for Malaysian kaolin clay is approximately 4; therefore,
after performing a holistic fit of the database, fsr can be expressed as Figure 11 Results of improved prediction method
below:
fsr =1 +  ((’v / su) / 4 – 1) (3)

where’v = soil effective stress (kN/m2); su = soil shear strength

(kN/m2);  = empirical factor = 0.4 (-). The value of the 0.4
provides the best fit to the experimental database.

5.8 Net extraction load, Quplift

The net extraction load Quplift is computed in the improved method
as:

Quplift = Qtop + Qbase + Weff (4)

Qtop = 0.25  D2 (Nc,top su,top fg,top) (MN) (5)

Figure 12 Comparing performance of reference method with

Qbase = 0.25  D2 (fsr Nc,base su,base fg,base fol –’H) (MN) (6)
updated formulation

60
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 45 No.4 December 2014 ISSN 0046-5828

Table 4 Predictions using updated parameters

Test Shear Shear Bearing Bearing Normalised Factor of Factor of Top Base Predicted Error
name strength strength factor at factor at strength strength operation resistance resistance breakout
gain at gain at top base ratio load
top base
- fg,top fg,base Nc,top Nc,base (- 'v/su fsr fol Qtop Qbase (MN) Qbreakout -
(-) (-) (-) ) (-) (-) (-) (MN) (MN) (%)
3.0D2.0Y 0.84 1.75 12.56 5.20 5.64 1.16 1.14 -2.84 -3.63 -6.47 5%
2.5D2.0Y 0.84 1.75 12.56 5.20 5.64 1.16 1.14 -2.33 -3.00 -5.33 -5%
2.0D2.0Y 0.84 1.75 12.56 5.20 5.50 1.15 1.14 -1.82 -2.38 -4.20 2%
1.5D2.0Y 0.84 1.75 12.56 5.20 5.50 1.15 1.14 -1.32 -1.76 -3.07 -7%
1.5D3.0Y 0.86 1.84 12.56 5.20 5.50 1.15 1.14 -1.35 -1.93 -3.28 -4%
1.5D1.0Y 0.81 1.66 12.56 5.20 5.50 1.15 1.14 -1.27 -1.59 -2.85 1%
1.5D0.5Y 0.77 1.54 12.56 5.20 5.50 1.15 1.14 -1.21 -1.36 -2.57 6%
1.5D0.0Y 0.60 1.00 12.56 5.20 5.50 1.15 1.14 -0.94 -0.35 -1.29 -18%
Nojet2 0.83 1.75 12.56 5.20 7.21 1.32 1.14 -6.26 -8.18 -14.44 4%
Nojet1 0.83 1.75 12.56 5.20 6.94 1.29 1.14 -6.54 -8.54 -15.09 6%
S1UEnJ 0.88 1.85 12.56 5.17 5.13 1.11 1.16 -32.79 -48.54 -81.33 1%
S2UEnJ 0.88 1.85 12.56 4.84 5.13 1.11 1.16 -22.20 -31.16 -53.36 14%
GS1 0.60 1.00 12.56 5.16 4.17 1.02 1.10 -10.23 -5.55 -15.78 -10%
GS2 0.68 1.25 12.56 5.20 4.17 1.02 1.10 -12.25 -11.23 -23.47 20%
GS3 0.74 1.45 12.56 5.20 4.17 1.02 1.10 -13.33 -15.43 -28.76 17%
GS4 0.78 1.58 12.56 5.18 4.17 1.02 1.10 -13.58 -17.55 -31.13 12%
GS5 0.82 1.69 12.56 5.20 4.17 1.02 1.10 -14.87 -20.59 -35.46 13%
GS6 0.86 1.80 12.56 5.20 4.17 1.02 1.10 -15.39 -22.68 -38.06 5%
D-01 0.81 1.68 12.45 4.97 3.90 0.99 1.00 -11.94 -13.53 -25.47 8%
D-02 0.81 1.68 12.56 5.20 3.93 0.99 1.00 -15.78 -18.19 -33.97 7%
D-03 0.81 1.68 12.56 5.20 3.95 1.00 1.00 -18.97 -21.15 -40.12 6%
C-03 0.60 1.00 12.56 5.20 3.95 0.99 1.00 -13.49 -5.28 -18.77 -7%
C-02 0.60 1.00 12.56 5.20 3.93 0.99 1.00 -11.67 -4.72 -16.38 -14%
C-01 0.60 1.00 12.56 5.16 3.93 0.99 1.00 -11.09 -4.38 -15.47 -11%

7. CONCLUSION 9. REFERENCES
A database of centrifuge tests on spudcan extraction in two different Bienen, B., Gaudin, C., and Cassidy, M.J. (2009) "The influence of
types of clay has been gathered to assess the performance of an pull-out load on the efficiency of jetting during spudcan
analytical method developed to predict the peak extraction extraction". Applied Ocean Research, Vol. 31, No. 3, pp 202-
resistance. The method proved to predict accurately the 211.
experimental results in Malaysian clay, but exhibited a significantly Finnie, I.M.S., and Randolph, M.F. (1994) "Punch-through and
lower performance for UWA kaolin clay with a mean difference of liquefaction induced failure of shallow foundations on
about 57%. calcareous sediments". Proc. Int. Conference on Behaviour of
A set of recommendations is proposed to update and improve the Offshore Structures, Boston, USA, pp 217-230.
prediction method. The recommendations relates to the factors Gaudin, C., Bienen, B. and Cassidy, M.J. (2011) "Investigation of
characterising the change in soil shear strength at the base and on the potential of bottom water jetting to ease spudcan
the top of the spudcan and two new factors considering the effect of extraction in soft clay". Géotechnique, Vol. 61, No. 112, pp
the operation load and strength ratio on spudcan extraction in clay. 1043-1054.
Additional details predicting when flow around occurs during Garnier, J., Gaudin, C., Springman, S.M., Culligan, P.J., Goodings,
installation have also been incorporated. The improved method D., Konig, D., Kutter, B., Phillips, R., Randolph, M.F., and
demonstrated a higher degree of accuracy with a mean difference Thorel, L. (2007) "Catalogue of scaling laws and similitude
reduced to 8% for both types of clay. questions in centrifuge modelling". Int. Journal of Physical
Modelling in Geotechnics, Vol. 7, Issue. 3, pp 1–23.
8. ACKNOWLEDGMENTS
Hossain, M.S., Randolph, M.F., Hu, Y. and White, D.J. (2006)
This work forms part of the activities of the Centre for Offshore "Cavity stability and bearing capacity of spudcan foundations
Foundation Systems (COFS), currently supported as a node of the on clay". Proc. Offshore Technology Conference, Houston,
Australian Research Council Centre of Excellence for Geotechnical USA, OTC 17770.
Science and Engineering and as a Centre of Excellence by the Houlsby, G.T. and Martin, C.M. (2003) "Undrained bearing
Lloyd's Register Foundation. Lloyd’s Register Foundation invests in capacity factors for conical footings on clay". Géotechnique,
science, engineering and technology for public benefit, worldwide. Vol. 53, No. 5, pp 513-520.
This research is supported by the Robert and Maude Gledden Osborne, J.J., Teh, K.L., Houlsby, G.T., Cassidy, M.J., Bienen, B.,
Postgraduate Research Scholarships and the Centre for Offshore and Leung, C.F. (2011) "InSafeJIP : Improved guidelines for
Foundation Systems (COFS). The fourth author is the recipient of an the prediction of geotechnical performance of spudcan
Australian Research Council (ARC) Postdoctoral Fellowship foundations during installation and removal of jack-up units".
(DP110101603). This support is gratefully acknowledged. Joint Industry-funded Project.

61
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 45 No.4 December 2014 ISSN 0046-5828

Kohan, O., Bienen, B., Cassidy, M.J., and Gaudin, C. (2013a)

"Centrifuge experiments to study extraction of a deeply
embedded spudcan using top jetting". Proc. 32nd
International Conference on Offshore Mechanics and Arctic
Engineering (OMAE), Nantes.
Kohan, O., Gaudin, C., Cassidy, M.J., and Bienen, B. (2013b)
"Spudcan extraction from deep embedment in soft clay".
Applied Ocean Research, Submitted in September 2013.
Kohan, O., Bienen, B., Gaudin, C., and Cassidy, M.J. (2014) "The
effect of water jetting on spudcan extraction from deep
embedment in soft clay". Ocean Engineering, Submitted in
January 2014 .
Merifield, R.S., Lyamin, A.V., Sloan, S.W. and Yu, H.S. (2003)
"Three-Dimensional Lower Bound Solutions for Stability of
Plate Anchors in Clay". Journal of Geotechnical and
Geoenvironmental Engineering, Vol. 129, No. 3, pp 243-253.
Purwana, O.A., Leung, C.F., Chow, Y.K., and Foo, K.S. (2005)
"Influence of base suction on extraction of jack-up spudcans".
Géotechnique, Vol. 55, No. 10, pp 741-753.
Purwana, O.A., Leung, C.F., Chow, Y.K., and Foo, K.S. (2006a)
"Breakout failure mechanism of jackup spudcan extraction".
Proc. 6th Int. Conf. of Physical Modelling in Geotechnics,
ICPMG06, Hong-Kong, Ng, Zhang & Wang (eds), pp 667-
672.
Purwana, O.A. (2006b) "Centrifuge model study on spudcan
extraction in soft clay". PhD Thesis, National University of
Singapore.
Purwana, O.A., Quah, M., Foo, K.S., Nowak, S., and Handidjaja, P.
(2009) "Leg Extraction / Pullout Resistance - Theoretical and
Practical Perspectives". In. Proc. 12th Jack up Conf., London.
Purwana, O.A., Krisdani, H., Zheng, X.Y., Quah, M., and Foo, K.S.
(2010) "An assessment of jack up spudcan extraction". Proc.
Int. Symp. on Frontiers in Offshore Geotechnics, Perth,
Australia, pp 679–684.
Purwana, O.A. (2010) "Personal correspondence".
Zhou, X.X., Chow, Y.K. and Leung, C.F. (2009) "Numerical
modelling of extraction of spudcans". Géotechnique, Vol. 59,
No. 1, pp 29-39.

62