See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/336242666

Advanced site-specific analysis of skirted spudcans in the view of North Sea

experiences

Conference Paper · September 2019

CITATIONS READS

0 159

11 authors, including:

Harun Kürşat Engin Hans Petter Jostad

Norwegian Geotechnical Institute Norwegian Geotechnical Institute
40 PUBLICATIONS   280 CITATIONS    142 PUBLICATIONS   1,642 CITATIONS   

SEE PROFILE SEE PROFILE

Marco D'Ignazio Nallathamby Sivasithamparam

Tampere University Norwegian Geotechnical Institute
32 PUBLICATIONS   178 CITATIONS    57 PUBLICATIONS   474 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Shallow CPT interpretation View project

NGI strategic project SP9: Behaviour of Sand under Partial Drainage and Offshore Foundation Design View project

All content following this page was uploaded by Harun Kürşat Engin on 08 October 2019.

The user has requested enhancement of the downloaded file.

ADVANCED SITE-SPECIFIC ANALYSIS OF SKIRTED SPUDCANS IN THE VIEW OF NORTH SEA
EXPERIENCES

H.K. Engin*, H.P. Jostad, M. D'Ignazio, N. Sivasithamparam, H.D.V. Khoa, K.H. Andersen, J. Johansson,
A.M. Kaynia and Ø. Torgersrud
NORWEGIAN GEOTECHNICAL INSTITUTE (NGI)

G. Yetginer
EQUINOR

H. Hofstede
GUSTOMSC
* corresponding author: [email protected]/[email protected]

ABSTRACT
Depending on the soil profiles and service conditions, site-specific assessment of jack-up platforms may require
detailed and advanced geotechnical foundation assessments. In these assessments, foundation fixity or global soil
spring stiffnesses play a significant role on dynamic behaviour and structural utilization of jack-up platforms,
especially at large water depths. In order to establish reliable values for this fixity, it is vital to consider the complex
non-linear behaviour of the different soil layers involved under combined average and cyclic loading. This paper
summarizes the challenges and possible solutions of site-specific foundation assessment of jack-up structures with
skirted spudcans, focusing on the estimation of monotonic and cyclic soil parameters, foundation penetration,
capacity, dynamic and quasi-static stiffnesses, damping, earthquake analyses, large deformation effects, rate
effects in the view of North Sea experiences.

KEY WORDS: Skirted spudcan, site-specific assessment, capacity, fixity, damping, advanced finite element
analyses.

INTRODUCTION
Jack-up rigs are commonly deployed as mobile units for a wide spectrum of offshore activities, ranging from
exploration and production of hydrocarbons to installation of wind turbines and their foundations. Site-specific
assessment of the jack-ups is required as recommended in ISO [1]. ISO [1] specifies requirements and guidance
for the site-specific assessment of independent leg jack‑up units, which should include both extreme events and
operational conditions. Proof of the robustness of the jack-up during the design storms makes it often necessary to
take moment fixity of the footings into account in order to reduce the critical moment in the lower leg guides,
reducing the inertia forces (i.e. reducing the natural period of the platform) and thus reducing the second-order
moment in the leg (P-Δ effect). Footings with skirts that penetrate into the soil generally improve the moment
fixity and bearing capacity compared to standard spudcans. However, from the assessment point of view, [2] has
shown that the design soil parameters and related capacity and stiffness of skirted spudcans should be carefully
considered as the simplified methodology described in ISO [1] is not always conservative, for instance for highly
overconsolidated clays.

This paper presents the challenges and possible solutions for the advanced site-specific foundation assessment of
skirted spudcans of jack-ups in the view of North Sea experiences in particular focusing on soil parameter
assessment and advanced finite element analyses.

ASSESSMENT OF SOIL PARAMETERS

Assessment of soil parameters is key to a reliable geotechnical assessment of jack-up foundations. For reducing
the uncertainties in the estimated foundation capacities and fixities, location-specific soil investigations are vital.
Indeed, certifying authorities require a site-specific assessment of the jack-ups to reduce uncertainties and assure
robustness of the jack-up structures [3]. In special cases, e.g. where the legs are situated in different soil units, it
may be beneficial to perform leg specific analyses. It may be relevant to consider a predominant loading direction
in addition.

This section summarizes different aspects of soil parameter assessment from monotonic to cyclic, highlighting
challenges and possible solutions. In addition, some background on the cyclic accumulation and cyclic contour
diagram concepts [4] is presented.

In-situ and lab measurements

Soil investigations are the key to geotechnical assessments. It is important to have a good coverage of the soil
volume affecting the foundation response with high quality and relevant field and laboratory measurements,
especially for advanced analyses. For sites with significant spatial variations in layering and soil properties, the
uncertainty levels may become significant and need to be carefully considered in the assessment. Laboratory tests
on undisturbed samples should be used to establish a reliable CPT interpretation framework, especially for low
permeable dilative soils, e.g. OC clays and (dense) silts.

In the presence of OC clays and (dense) silts, it should be considered that the stress-strain behaviour of these soils
is characterized by dilatancy. Further, these soils generally show a fairly stiff response during monotonic loading,
while the behaviour becomes softer during cyclic loading. The behaviour is strongly affected by the content of
fines, water content and stress history [4].

Cyclic laboratory tests can be costly but are quite valuable since they may reduce the uncertainties, with the
condition that they are performed on high-quality samples and following anticipated stress paths (as illustrated in
Figure 1). The location and number of samples should be carefully selected. The missing info can be estimated by
employing published databases [4].

Figure 1 Stress paths beneath a foundation under combined HVM-loading [4]

Monotonic parameters
Finite element analyses have shown that the behaviour may be strongly dominated by the weakest zone near skirt
tip level, as opposed to the concept assuming an average strength in a zone down to the spudcan diameter. As a
result, lack of data at these shallow depths, especially for OC clays and silts, brings the challenge of extrapolating
design shear strength profiles to skirt tip level and seabed. For example, the cone factor 𝑁 used for assessing the
undrained shear strength of the soil can be quite high for high OCR clays, which is generally difficult to estimate.
In dilative soils, the laboratory strength is generally selected at a threshold strain level (i.e., 10% axial strain in
triaxial or 15% shear strain in direct simple shear tests), if the peak shear strength has not been reached before.
However, this threshold strain level may be far lower than the strain level at which the maximum cone resistance
is reached during CPT testing. The mechanism is schematized in Figure 2. Therefore, the selection of 𝑁 should
take this aspect into account. For sites where the OCR varies significantly with depth, the SHANSEP approach [5]
can be employed to combine the OCR data and shear strength data for a more reliable 𝑁 assessment and hence,
more reliable shear strength profiles, as suggested by [6]. This will result in a variable 𝑁 with depth, as shown
in Figure 3, compared to the more classical constant-𝑁 (with 𝑁 15 20) approach which does not take the
stress-history of the deposit into account. Assuming a constant 𝑁 value may overestimate the design shear
strength at shallow depths where OCR may be high.

Figure 2 Large shear strains around a CPT as an explanation for the high 𝑁 factors for dilative soils.

Figure 3 Example characteristic static triaxial shear strength assessment.

Similarly, standard assessment of OC sands based on CPT generally results in overestimated relative densities.
Hence, one should consider these issues in the assessment, as the relative density is linked to the friction angle of
the sand and hence to the shear strength [4].

Establishing characteristic design profiles can be challenging considering statistics since quite limited laboratory
data is available in the majority of the site investigations. Generally, engineering judgement is required for
establishing the design profiles.

In advanced foundation analyses, it is important to assess the effect of stress-path dependency, which is generally
referred to as strength anisotropy. However, soil investigations generally lack sufficient test data at representative
locations and depths. It is then common to perform different tests (e.g. DSS, CAUC and CAUE) on samples from
different boreholes and depths and combine them to characterize a given soil unit.
Since jack-ups are normally preloaded during installation, the effect of preloading on the soil strength and stiffness
parameters should be considered. In general, it is often assumed that preloading has a positive effect since after
unloading to the operational load, the material is in an unloading/reloading state. This should imply a stiffer
material response compared to a virgin loading state. However, this is not true for highly overconsolidated soil as
frequently found in North Sea, since the material is already in an unloading/reloading state. In addition, in order
to account for any positive effect of preloading, the soil needs to be drained under the increased vertical stresses.

Permeability
Hydraulic conductivity, generally referred to as permeability, is another crucial parameter for the assessment of
silt and sand behaviour, i.e. drained, partially drained or undrained response at different operational conditions.
Permeability is generally evaluated from oedometer tests. However, when limited or no data is available, the
assessment should be based on basic soil parameters (i.e., water content, void ratio, grain size) and databases [7],
as illustrated in Figure 4. The drawback of such a procedure is that the determination becomes empirical and may
inherit a high level of uncertainty. The foundation response can be quite sensitive to the estimated permeability
values.

Figure 4 Estimation of permeability [7]

Similarly, stiffness parameters (i.e., constrained modulus), which can be estimated from oedometer tests as well
as empirical correlations, are important for the assessment of drainage conditions together with permeability.

Foundation loads
The foundation loads (Figure 5) are obtained from dynamic and quasi-static structural analyses with input of
foundation stiffnesses. Therefore, the foundation loads and stiffnesses need to be calculated in an iterative manner.

One of the main challenges is to define the average and cyclic load components (and corresponding cyclic/average
ratios for the assessment of cyclic soil properties). One can define the average loads as the average of the peak-
through values.

The assessment of cyclic stress-strain curves, as for instance shown in [2], requires the cyclic/average load ratio
as input. This ratio is not unique for a given storm load history. It is load level as well as load component (V, H,
and M) dependent. One may therefore select the ratio from the global overturning moment at seabed which dictates
the foundation stiffness.

 
Figure 5 Sketch of a simplified structural model of jack-up with different global loads and foundation reactions.

 
Figure 6 Total foundation load history decomposed into average and cyclic components, idealization of the load history and
calculated equivalent load cycles.

Strain or pore-pressure accumulation procedures should be employed to estimate an equivalent number of constant
load cycles (𝑁 ) giving the same amount of cyclic degradation as the irregular storm load history. For this
purpose, the load histories need to be simplified into load parcels, which in many cases are sorted from the smallest
to the largest amplitude with the aim to obtain the largest cyclic degradation for the maximum loads (Figure 6).
How to determine the idealized load history is discussed in [8].
The basis for obtaining the stress-strain curves are the cyclic contour diagrams, which contain the average and
cyclic shear strains for a given number of cycles of constant cyclic shear stress amplitudes and average shear
stresses. The effect of preloading can be taken into account when performing the cyclic tests as a basis for the
contour diagrams.

Determination of cyclic parameters

As mentioned briefly in the previous section, strain or pore-pressure accumulation procedures should be employed
to estimate an equivalent number of load cycles (𝑁 ) of the maximum loads that gives the same cyclic effects
(i.e. reduced shear strength or accumulated pore pressure) as the entire storm load history. The method is applied
to each layer to estimate layer-specific cyclic degradation and eventually utilize the most appropriate cyclic
contour diagram [4]. Figure 7 illustrates the effective stress paths in undrained test of a contractive soil and
definition of parameters as input to cyclic contours diagrams. Figure 8 presents an example of 3D representation
of cyclic and average shear strains as functions of average and cyclic shear stresses and the number of cycles [4].

Figure 7 Effective stress paths in undrained test of a contractive soil and definition of parameters as input to cyclic contours
diagrams [4].

Figure 8 Example 3D representation of cyclic and average shear strains as functions of average and cyclic shear stresses
and the number of cycles – Normally consolidated Drammen Clay [4].
The accumulation procedures can further account for partial drainage. However, the sorting of load parcels, i.e.
from smallest to the highest amplitude which dictates the accumulation, is generally conservative. In reality, the
largest loads do not come immediately after each other, allowing for partial drainage between different large load
cycles, especially in sandy and silty soils.

The accumulation procedures should account for both horizontal and vertical drainage when relevant. Depending
on the project demand and complexity, the Partially Drained Cyclic Accumulation Model (PDCAM) can be
utilized for accounting for partial drainage throughout the whole soil domain during cyclic loading.

Ideally, one should construct contour diagrams for each soil layer for the specific location of the spudcans.
However, construction of the contour diagrams requires a significant number of cyclic tests. Therefore, databases
for different soil types are often utilized with adjustments for the site/location-specific properties for establishing
contour diagrams. If available and reliable, it is recommended to use the site/location-specific cyclic tests to adjust
the contour diagrams originating from databases.

The cyclic tests incorporate the rate effects, which increase the cyclic shear strength and stiffness for a low number
of cycles compared to standard monotonic tests. Hence, the negative effect of cyclic degradation might be
compensated and even overruled by rate effects for low 𝑁 .

As for the monotonic behaviour, an important aspect of cyclic soil behaviour is the cyclic strength anisotropy, i.e.
stress path dependent cyclic shear strength as illustrated in Figure 1.

Establishing input for capacity and stiffness (FE) analyses

Finite element (FE) analyses of the spudcan foundations require representative soil behaviour, which can be
described by different constitutive models. The methodology described in [2] and in this paper utilizes two sets of
input: cyclic and total (average+cyclic) shear stress-strain curves for calculation of foundation stiffnesses as input
to dynamic and quasi-static structural analyses.

Figure 9 Sketch of the utilization of contour diagram for the extraction of stress-strain curves for jack-ups with a given
weight/still water load (SWL)

Figure 9 illustrates the use of cyclic contour diagrams, for the assessment of shear strength versus strain curves as
input to the calculation of foundation capacity and stiffnesses. The procedure combines the effect of cyclic soil
response with foundation loads as well as cyclic to average load ratios. The still water load determines the initial
mobilization. Hence the cyclic stress-strain curves can be sensitive to SWL load level.
 Load

<

<

<

Figure 10 Sketch of vertical load-displacement response, accumulated and cyclic displacements, and determination of secant
stiffnesses.

One important challenge is the summation of cyclic and average responses. The example in Figure 10 is for the
windward leg: one should be aware of the accumulated average displacements when calculating total stiffnesses.
In other words, the cyclic shear stresses (which in this content always are positive values) from the cyclic loads
should not reduce the (average) shear stresses from the average loads. In order to account for this effect, the cyclic
shear strains (amplitude) always should be added to the average shear strain when calculating the total shear strain
(sum of average and cyclic shear strain) and thus be consistent with the data in the cyclic contour diagrams.
Therefore, the absolute value of the cyclic vertical load should be added to the average vertical leg load when
calculating the rotational stiffness of the windward leg (see Figure 10). Alternatively, the analyses could be divided
into calculations of the cyclic and average response, and adding these together.

PENETRATION ANALYSES
Simplified bearing capacity
Industry practice utilizes mostly the simplified bearing capacity methods for the calculation of possible spudcan
leg penetration assessments. The methods become as advanced as accounting for different layer interactions,
backflow, etc., e.g. [9]. Nevertheless, the methods have quite many limitations due to their empirical nature. For
example, these models cannot account for the effects of geometrical changes due to large deformations.

Lagrangian (small deformation) FEAs

One can consider the effect of geometrical changes, by employing simplified FEAs, e.g. [10] and [11]. These
methods have their limitations and not very well suited for deep penetrations as they cannot capture complex
mechanisms (e.g. backflow).

Large deformation FEAs

Spudcan penetration often involves large deformation in surrounding soil as well as the possible backflow soil.
Although having advantages of both computational simplicity and efficiency, the conventional small-strain
Lagrangian approach and the Updated Lagrangian (i.e. updated mesh) approach used in the finite element method
may suffer from serious numerical difficulties when extreme mesh distortion occurs. Several advanced modelling
techniques including Material Point Method (MPM), Smooth Particle Hydrodynamics (SPH), Particle Finite
Element Method (PFEM), Remeshing and Interpolating Technique with Small Strain (RITSS), Arbitrary
Lagrangian-Eulerian (ALE), etc. have been developed to overcome the numerical issues. Among them, the
Coupled Eulerian-Lagrangian (CEL) technique has been widely used at NGI and been proven as a suitable and
efficient modelling approach for capturing complex mechanisms and geometric nonlinearity effect during spudcan
penetrating through layered soils. For example, [12], [13], [14] successfully applied the CEL and the ALE methods
to evaluate the installation effects and compute the bearing resistance of offshore anchors and foundations.
Through a series of back-analyses of different centrifuge tests of spudcan penetration, [9] and [15] found out and
discussed main advantages of using the CEL method compared to some semi-empirical methods recommended in
industry guidelines and the small strain finite element method with the Press-Replace (PR) technique [11]. Several
large deformation FE analyses using the CEL were also performed by [15] to back-calculate results of centrifuge
tests of unstiffened/stiffened suction anchors as well as to evaluate potential interaction between spudcan and
neighbouring subsea template during jack-up installation. Very good agreement between the numerical predictions
and the measurements was observed.

Recently, [16] applied the CEL technique in Abaqus to investigate the interaction of spudcan and soil in different
sloping angles of the seabed and compared with the existing centrifuge tests. Reasonably good agreement was
found by the authors.

A good example of challenging soil conditions with silty layer can be seen in Figure 11. Considering dilation
indicates significantly lower penetration depth with high potential of punch-through.

Figure 11 Summary of large deformation Abaqus CEL analyses of spudcan penetration.

Jack-up analyses require foundation capacity checks with load and material factors. The standard approach is
employing ISO [1]. At NGI, axisymmetric analyses (with higher-order elements) are used as a reference for
calibrating 3D FEAs, which inherit overshoot. A robust and cost-effective 2D model with side shear (HVMCap /
BIFURC2D) can be calibrated for vertical capacity, e.g. using FEAs employing higher-order elements, i.e. Plaxis
axisymmetric analyses using 15-noded triangular elements. The assumption is that the side shear factors are valid
for combined loading (i.e. VHM). Cyclic soil shear strength values are used for obtaining capacity envelopes.

Several analyses at different sites and locations indicate that the failure modes can be complex (see Figure 12) and
generally quite sensitive to shear strength profile just below the skirt tip level. In this cases the calculated capacity
is found to be significatly lower than obtained from cassical bearing capacity equations using an average strength
over a depth equal to the diameter of the footing.
 Figure 12 Example failure mechanisms (contours of incremental shear strains) for two different soil profiles

FOUNDATION STIFFNESS
Foundation stiffnesses (fixity) or global soil spring stiffnesses play an important role on dynamic behaviour and
structural utilization of jack-up platforms, especially at large water depths. In order to establish accurate values for
this fixity, one needs to account for the complex non-linear behaviour of the different soil layers involved under
combined average and cyclic loading. The procedure for establishing non-linear stress-strain relationships to be
used as input to finite element analyses of jack-up footings together with challenges associated is presented in the
previous section. The calculated non-linear load-displacement relationships (foundation stiffnesses) of the
individual footings are divided into cyclic and total (average plus cyclic) components to be used as input to the
dynamic and quasi-static structural analyses of the jack-up, respectively.

Since the stress-strain curves used in the analyses are based on results from cyclic tests, and therefore only are
valid at the actual stress state after an equivalent number of cycles, it cannot be used to follow the actual load
history as shown in Figure 13. Instead, the loads are applied monotonically from zero to the actual loads and only
the results at the end of the analysis are used.
M / V / H

M / / 
/ 

Secant rot. 
stiffness curve
/ 

M / V / H
Figure 13 Procedure for calculation of foundation secant stiffness curves

The aforementioned secant stiffness curves are sensitive to different load components, i.e. there is an effect of
coupling between for instance the horizontal load and the overturning moment (see Figure 15). The calculated
stiffnesses can be best represented as matrices with several off-diagonal coupling terms. However, as sketched in
Figure 14, at a load reference point (LRP) where negligible horizontal displacements evolve with the applied
moment, stiffnesses can be represented as vectors. Although it gives a better understanding of the foundation
response, it is less practical to use the LRP due to its load level dependency. Hence, in practice, one can ignore the
coupling terms as long as the load levels and ratios, and the attachment point are kept same, and then perform
several stiffness analyses to cover anticipated range of loads to provide good coverage of stiffness for more
efficient iteration with the structural analyses.

Figure 14 a) Definition of load and displacement vectors, b) Physical interpretation of the load reference point LRP. [17]

Figure 15 Effect of horizontal load on secant stiffness response

Figure 16 presents the cross-section of an example three-dimensional (3D) finite element (FE) model of the NGI
in-house FE program INFIDEL. The INFIDEL program uses near-field and far-field elements to model the soil
volume and a cylindrical boundary separate them. Near-field elements are ring elements and their upper and lower
faces are horizontal, while their inner and outer faces vertical circular cylinders. The elements are interconnected
along the four corner node circles. A column of thin elements was used to model the interface behaviour along the
outer skirt. Due to inner skirts that divide the soil into several compartments, a zone of very stiff linear elastic
material up to inner skirt tip level (1 m) can be assumed (assuming complete skirt penetration with a full contact
with base).
 Figure 16 Example finite element mesh of INFIDEL used for stiffness calculations

Figure 17 presents the finite element calculation results (shear stress mobilization ratios) of Plaxis 3D for two
different layering and load cases. These type of plots are very useful for the assessment of the contribution of the
soil reaction and deformations of different layers.

Figure 17 Example finite element calculation results (shear mobilization ratios) of Plaxis 3D for two different layering and
load cases.
SOIL AND FOUNDATION DAMPING
Soil material damping reduces the resonant response of jack-up platforms and the corresponding demands in the
leg's truss members and the foundation loads. To optimize the design, hysteretic foundation damping, for use in
the structural analysis, can be computed with different FE-tools (e.g. INFIDEL, or Comsol/Plaxis). Due to the low
vibration frequency of typical wave loads acting on jack-up foundations, soil radiation damping does not contribute
to foundation damping, while it may contribute to some extent to foundation damping for earthquake loading.

To compute load-dependent hysteretic foundation damping, hysteretic soil damping versus strain curves are
needed for each soil layer (see Figure 18a) in addition to stress-strain curves, where the global hysteretic
foundation damping is computed as,
𝐸
𝐷
4𝜋𝐸

the ratio of the integration of hysteretic, 𝐸 , and elastic energy, 𝐸 , in all soil elements.

Since the soil strain diminishes with distance from the foundation, the largest contribution to the foundation
damping comes from the soil in the vicinity of the foundation. As an example, Figure 18b shows the distribution
of soil material damping around a skirted spudcan foundation during storm loading. The soil slightly larger than a
spherical zone following spudcan diameter contributes the majority of the foundation damping, however the size
of this zone will vary with the type of foundation and loading (moment, horizontal, vertical), and soil properties.

20.0 Clay Ia
18.0
16.0 Clay I a
14.0
12.0
D [%]

10.0
8.0
6.0
4.0
2.0
0.0
0.0001 0.001 0.01 0.1 1 10
Shear strain %

Figure 18 a) Example of soil damping curve for a clayey material. b) Example of calculated soil material damping around a
spudcan for a ULS load case.

The soil damping and thus also foundation damping is dependent on the soil strain and will vary with the loads
acting on the foundation. Site-specific strain-dependent soil damping can be evaluated based on laboratory test.
To establish a complete damping curve over a wide range of soil strain a combination of standard laboratory tests
such as Resonant Column test (RC) and DSS/TX is used. There are large amount of laboratory-based soil damping
curves in the literature see e.g. [18],[19],[20],[21] and references therein in case of lack of site-specific soil samples
or in early phases of projects. In addition to soil shear strain other parameters that have an important influence on
the material damping is the plasticity index, the effective confining pressure, and over consolidation ratio.

Some researchers have observed a rate dependency of the damping [22] with an increase in damping value with a
factor 2 per log cycle strain rate. The damping's rate dependency should be considered when performing and
evaluating laboratory test and when computing foundation damping. Further research is needed to understand what
mechanisms in the soil are causing the damping, under which conditions do they occur and are the mechanisms
the same in all types of soils.

For offshore structures subjected to wave loading both average and cyclic loads are important for the foundation
response as described earlier. To evaluate properly the soil damping in laboratory tests only the soil damping due
to cyclic loading is of interest. To remove the apparent damping due to permanent strains observed in laboratory
experiments different approaches have been proposed and evaluated by [23].
The foundation damping ratios predicted for spudcan foundations for jack-up for storm loading are often large
(𝐷 10 15%). An example of foundation damping versus footing moment load for a typical spudcan
foundation is shown in Figure 19 for the windward, leeward and neutral leg. The foundation damping increases
quickly with increasing moment load and increasing vertical load.

16.0

14.0

12.0

10.0
V = 29.5 MN (WW1 ‐ ULS)
V = 29.5 MN (WW1 ‐ ULS)
Damping [%]

8.0
V = 29.5 MN (WW1 ‐ ULS)
6.0 V = 13.1 MN (WW2 ‐ ULS)
V = 13.1 MN (WW2 ‐ ULS)
4.0 V = 13.1 MN (WW2 ‐ ULS)
V = 0 MN (Neutral ‐ ULS)
2.0 V = 0 MN (Neutral ‐ ULS)
V = 0 MN (Neutral ‐ ULS)
0.0
0 50 100 150 200 250 300 350 400
Moment [MNm]

Figure 19 Example of damping versus moment curve for different vertical and horizontal loads for windward and leeward
legs.

1.0 18
0.9 16
0.8 14
Normalized Stiffness

0.7 G/Gmax
12
Damping [%]

0.6 Foundation Stiffness
10
0.5 Soil Damping
8
0.4 Foundation Damping
0.3 6
0.2 4
0.1 2
0.0 0
0.000001 0.00001 0.0001 0.001 0.01
Strain or Rotation

Figure 20 Normalized soil and foundation stiffness, and soil and foundation damping, versus shear strain and rotation for a
constant soil profile. 

The hysteretic foundation damping shown in Figure 19 cannot be used directly in the structural time-domain
analysis and the damping could be converted to dashpot placed at the mudline [24]. However, such dashpot will
be frequency-dependent, and if calibrated for the frequency of the fundamental vibration mode it will likely
overestimate the foundation damping for higher vibration modes. An improved alternative to dashpot is a
kinematic hardening model with modified foundation moment-rotation response [25]. Further improved models
for accounting for foundation damping are so-called macro-element models [26][27][28] and can be calibrated for
different loading conditions and load-dependent damping.

Recent analyses with a constant soil profile indicate there may be strong similarities between rotational foundation
stiffness and damping versus rotation and soil stiffness and damping versus shear strain curves as shown in Figure
20. Further investigation may allow for a simple procedure to establish foundation stiffness and damping which
can be useful in early phases of projects when project-specific stress-strain curves are not available.

EARTHQUAKE ANALYSES
Earthquake analyses of jack-ups are often carried out using the mode superposition technique and response spectra
of the design earthquake. For such analyses, one traditionally needs the foundation stiffnesses for the spudcans,
sometimes referred to as soil springs. These springs are often computed using equivalent linear methods with soil
moduli representing the level of soil nonlinearity during earthquake shaking.

A potentially problematic issue related to earthquake loading is soil liquefaction. Liquefaction is generally not a
design issue in the North Sea. However, considering the importance of liquefaction in seismic regions and major
impact on spudcan foundations, a brief account of it is included here.

Over the past few decades, laboratory tests (e.g. [29]) and CPT/SPT based liquefaction triggering procedures (e.g.
[30]) have been conducted to predict the liquefaction potential and the possible consequences of it on structures.
Despite extensive research performed on liquefaction, the empirical methods based on CPT and SPT are still the
most commonly used because they are based on the calibration of these methods against observed field data. In
the empirical methods, the normalized cyclic shear strength, often referred to as cyclic resistance ratio CRR, is
determined for the measured value of CPT/SPT and the earthquake magnitude. The CRR is then compared with
the earthquake-induced cyclic shear stress normalized by the effective vertical stress, denoted as the cyclic stress
ratio CSR.

While this procedure is commonly used for assessment of liquefaction in the free field, it can be extended to the
conditions under the structure by including the additional shear stresses from the structure's shaking in the CSR
and increasing the CRR for the additional overburden pressure from the structure. In cases where the structural
response to earthquake shaking is strong (like low to medium-rise buildings and GBSs), the increase in CSR
outweighs the increase in CRR, which results in a more critical condition under the structure. The situation is
opposite in structures with long natural periods (such as foundations of large bridges). Jack-ups often have long
natural periods; therefore, the liquefaction resistance under the spudcans is higher than the resistance outside the
footprint of the spudcans, and as a result, there is a possibility that in sites with liquefaction potential, the spudcans
have sufficient bearing capacity. A more reliable approach to assess the performance of complex structures, such
as jack-ups, in liquefiable soils is through advanced numerical FE simulations. A number of constitutive models
exist for liquefaction (e.g. [31][32][33]) which have been implemented in FE codes and have been applied to
various design cases (e.g. [34][35][36]). These models require diligent parameter calibration and extensive
sensitivity analyses to capture the range of possible responses.
ADVANCED FINITE ELEMENT ANALYSES
Estimating cyclic accumulation effects with undrained layers/boundaries
Low permeable soils that behave almost undrained within several hours subjected to cyclic and average loadings
will generally reduce the shear stiffness and the undrained shear strength of the soil. During the last 40 years, NGI
has developed procedures for modelling behaviour of soil subjected to combined average and cyclic loading.
Among them, the model called UDCAM (Undrained Cyclic Accumulation Model) is a finite element-based
calculation procedure that accounts for the effect of cyclic loading of soils under undrained conditions [37] using
the strain accumulation method [38]. Instead of analysing the cyclic load history in the time domain (implicit
method), it considers the behaviour during application of the loads in so-called load parcels of constant average
and cyclic load amplitudes (explicit method). The model then finds the equivalent number of cycles (𝑁 ), the
reduced cyclic stiffnesses and accumulated permanent strains and cyclic shear strength.

Estimating cyclic accumulation effects with partially drained layers/boundaries

Simplified pore pressure accumulation analyses cannot handle the dissipation of pore pressure for varying
permeability in different directions. In order to account for the partial drainage in all directions, through other
layers, the pore pressure accumulation can be done using FEAs with advanced constitutive models, e.g. the
Partially Drained Cyclic Accumulation Model (PDCAM) [39], which calculates the accumulated pore pressure
and the corresponding equivalent number of undrained cycles (𝑁 ) under cyclic loading. PDCAM does not predict
the behaviour within a cycle but predicts shear strain amplitudes and development of strains during a cyclic load
history. PDCAM is implemented in the NGI in-house FE program BIFURC.

Figure 21 presents an example 2D FE model as an axisymmetric model of a 22 m diameter spudcan with 2 m skirt
length. Drainage boundaries are specified at the seabed and the vertical outer boundary. No drainage through the
spudcan is allowed. PDCAM requires average and cyclic shear strain contour diagrams and pore pressure contour
diagrams.

Figure 21 Finite element model for PDCAM analysis to obtain 𝑁

A selected North Sea storm load history which includes 18-hours build-up and 6 hours peak period is applied as
load parcels to obtain 𝑁 at failure. Figure 22 presents the estimated equivalent number of cycles (𝑁 ) at failure
in the soil in a fully coupled pore pressure accumulation and dissipation analysis. PDCAM predicts much lower
degradation of soil (𝑁 10) in contrast to a simplified procedure considering only radial pore pressure
dissipation (𝑁 20).
  
Figure 22 Distribution of the equivalent number of cycles, 𝑁 , at the end of applied storm history scaled to cause failure.

Incorporating large deformation (geometrical change) effects

Large deformation finite element analyses (LDFEAs) e.g. Plaxis Updated Mesh, Abaqus CEL, ALE, can be used
to estimate displacements of the spudcan foundation during ULS/ALS loading by accounting for geometrical
changes. Figure 23 illustrates the example of expanding capacity envelope due to large vertical displacements
(penetration) of the spudcan, which can be predicted with LDFEAs.

spudcan
seabed

 
Figure 23 Illustration of expansion of capacity envelope due to large vertical displacements of spudcan

One can also consider the geometrical effects on foundation stiffnesses. Figure 24 compares the extent of shear
mobilization zones: noting that the lower the mobilization the higher the foundation stiffness. Depending on the
soil profile, one may obtain higher foundation stiffness using LDFEA (e.g. compression of the weakest soil beneath
the skirt tip level).
 
  

 
Figure 24 Contour plots of the degree of mobilisation of undrained shear strengths in soils for different load cases.
Calculated results of small strain (SS) FEA (left) and Updated Lagrangian (UL) FEA (right).

RATE EFFECTS
The cyclic laboratory tests already incorporate the rate effects if the tests are performed with the same load
frequency as the considered storm history. Storm loads can contain asymmetric peak loads, resulting in a sudden
change in average load, which can be faster than the incorporated rate effects in the reference lab tests (Figure 25).
Hence, one may utilize the additional rate effects if needed, considering (conservatively) the duration of equivalent
load cycles and database of [40] given in Figure 26. Figure 27 illustrates the simplified revision of a contour
diagram for an estimated increase of average shear strength to account for additional rate effects.

 
Figure 25 Example storm load history with asymmetric peaks.
  
Figure 26 Rate effects and estimation from database [40]

1)

2)

3)

 
Figure 27 Illustration of modification of contour diagram due to (additional) rate effects caused by the change in average
strains

DISCUSSIONS & CONCLUSIONS

Advanced foundation analyses of jack-ups provide valuable information in reducing the uncertainties of the
foundation response taking into account stress history, cyclic behaviour, rate effects large deformation effects and
complex soil layering. This paper highlights the important aspects and challenges of the advanced site-specific
analysis of jack-ups with skirted spudcans in the view of North Sea experiences, with particular focus on the
following aspects:
o Structural aspects:
o The simplified methodology described in ISO [1] may result in underestimation of the structural
utilisations.
o Soil parameters:
o The design strength profiles should be established preferably from in-situ (i.e., CPT) and laboratory
tests on adjacent boreholes (BHs). A proper CPT calibration allows extrapolating soil properties
between discrete sampling depths. In overconsolidated clays and silts, the selection of the cone factor
𝑁 should account for the OCR and its variation with depth. The selection of a constant 𝑁 may result
in unconservative design shear strength, especially at shallow depths that affect the spudcan capacity.
 o Soil stiffness and permeability are crucial to evaluate whether the foundation behaviour is governed by
drained, partially drained or undrained conditions. Relevant parameters should be determined from
laboratory oedometer tests or from databases when site-specific data is limited or unavailable.
o Cyclic properties should be preferably determined from site/layer-specific laboratory tests. In the
absence of sufficient site-specific data, cyclic contour diagrams can be selected from databases on
similar soils.
o Stress-path dependency of shear strength (i.e., anisotropy) should be investigated from triaxial and DSS
tests to perform advanced foundation analyses.
o Cyclic degradation of each soil layer, expressed in terms of the equivalent number of cycles (𝑁 )
resulting from a given storm load composition, can be determined according to a simplified strain/pore
pressure accumulation procedure. This procedure may be conservative unless it accounts for the actual
3D drainage path. Advanced FE accumulation procedure using PDCAM or UDCAM approaches
consider the interaction between the soil layers in the domain and is likely to result in lower Neq values
compared to a simplified approach.
o Loads:
o As loads and stiffness are dependent on each other, an iterative procedure is generally required.
o Cyclic degradation due to storm load history should be considered by evaluation of dynamic storm
loading from the jack-up onto the spudcan foundation by means of load parcels.
o Design storm loads from the jack-up on the foundation needs to be subdivided into what can be
considered as average and cyclic components. This is not a straightforward task considering the fact
that the foundation rotational stiffnesses are strongly dependent on the loads.
o Foundation stiffness and capacities:
o It is important to distinguish between cyclic foundation stiffnesses used in a dynamic structural analysis
and the total foundation stiffnesses used to check the integrity of the jack-up in a quasi-static analysis.
o The non-linear foundation stiffness depends on the soil type and loading condition.
o The moment fixity under the peak loads may be highly affected by cyclic degradation and strain
accumulation.
o Cyclic foundation stiffness and damping curves for use in the structural dynamic analysis are calculated
based on results from cyclic laboratory tests.
o The design soil parameters and related capacity and stiffness should be carefully considered as there is
no guarantee that the fixities calculated by standard ISO methods are always conservative for these
skirted spudcan foundations, for instance for highly overconsolidated clays.
o Large deformation effects can be employed to improve capacity and fixity in cases with increasing
shear strength profile. This also means that the effect of large deformation may be important to account
for in cases with decreasing shear strength with depth since the capacity and stiffness may be reduced.

ACKNOWLEDGEMENTS
The support of the Norwegian Geotechnical Institute and GustoMSC are greatly acknowledged.

REFERENCES
[1] International organisation for standardisation: Petroleum and offshore gas industries – site specific assessment
of mobile offshore units – part 1: Jack-ups (ISO 19905-1:2016), Geneve, 2016.
[2] Jostad HP, Torgersrud Ø, Engin HK, Hofstede H - A FE procedure for calculation of fixity of jack-up
foundations with skirts using cyclic strain contour diagrams, City university Jack-up conference, 2015, Sept
15-16.
[3] DNVGL-RP-C212: Offshore soil mechanics and geotechnical engineering, Edition August 2017,
https://rules.dnvgl.com/docs/pdf/DNVGL/RP/2017-08/DNVGL-RP-C212.pdf
[4] Andersen KH - Cyclic soil parameters for offshore foundation design. The 3rd ISSMGE McClelland Lecture.
Frontiers in Offshore Geotechnics III, ISFOG'2015, Meyer (Ed). Taylor & Francis Group, London, ISBN:
978-1-138-02848-7. Proc., p5-82. Revised version in: http://www.issmge.org/committees/technical-
committees/ applications/offshore and click on “Additional Information”.
[5] Ladd CC and Foott R - New design procedure for stability of soft clays. Journal of the Geotechnical
Engineering Division, ASCE, 100(7), 1974. p 763–786.
[6] Paniagua P, D'Ignazio M, L'Heureux JS, Lunne T, Karlsrud K - CPTU correlations for Norwegian clays: an
update, 2019. AIMS Geosciences 5.p 82–103. doi: 10.3934/geosci.2019.2.82
[7] Andersen KH and Schjetne K - Database of friction angles of sand and consolidation characteristics of sand,
silt and clay, Journal of Geotechnical and Geoenvironmental Engineering, ASCE, Vol. 139, 2013. p 1140-
1155.
[8] Norén-Cosgriff K, Jostad HP, Madshus C - Idealized load composition for determination of cyclic undrained
degradation of soils. In: V. Meyer (Ed.), Frontiers in Offshore Geotechnics III (ISFOG), Oslo, Norway, June
2015. Boca Raton: CRC Press. Chapter 152, p 1097-1102.
[9] Zhang Y, Khoa DVH, Meyer V, Cassidy M - Jack-up spudcan penetration analysis: Review of semi-analytical
and numerical methods, 3rd International Symposium on Frontiers in Offshore Geotechnics (ISFOG), Oslo,
Norway, June 2015.
[10] Andersen KH, Andresen L, Jostad HP, Cluckey EC - Effect of skirted-tip geometry on set-up outside suction
anchors in soft clay. In Proc. 23rd Int. Cof. on Offshore Mechanics and Arctic Eng., Vol.1, 2004. P 1035-
1044, Vancouver, Canada.
[11] Engin HK, Brinkgreve RBJ, van Tol AF - Simplified Numerical Modelling of Pile Penetration – the Press-
Replace technique, Int. J. An. and Num. Methods in Geomechanics, 2013.
[12] Andresen L, Khoa HDV - LDFE analysis of installation effects for offshore anchors and foundations. In:
Proceedings of the International Conference on Installation Effects in Geotechnical Engineering, Rotterdam,
2013. 162-168.
[13] Khoa HDV - Large deformation finite element analysis of spudcan penetration in layered soils. Third
International Symposium on Computational Geomechanics, Krakow, Poland, 2013. p 570-585.
[14] Khoa HDV - Numerical simulation of spudcan penetration using Coupled Eulerian-Lagrangian method, The
14th International Conference of the International Association for Computer Methods and Advances in
Geomechanics (14IACMAG), Kyoto, Japan, 2014.
[15] Khoa DVH, Jostad HP - Application of coupled Eulerian-Lagrangian method to large deformation analyses
of offshore foundations and suction anchors. Int. J. Offshore Polar Eng. 26(3), 2016. p 304–314
[16] Shin Y, Khoa HDV, Choi JC, Park HJ, Kim JH, Kim DS, Han JT, Lee SW, Choi J – Spudcan penetration
behaviour on a sloping seabed by numerical analysis and centrifuge modelling. In: Proc. 28th Int. Offshore
and Polar Eng'ng Conf. (eds J. S. Chung, B. S. Hyun, D. Matskevitch and A. M. Wang), Sapporo, Japan:
ISOPE: 2018. p 576–581.
[17] Skau KS, Grimstad G, Page AM, Eiksund GR, Jostad HP - A macro-element for integrated time domain
analyses representing bucket foundations for offshore wind turbines, Marine Structures, 59, 158-178.
doi:10.1016/j.marstruc.2018.01.01.
[18] Seed HB, Idriss IM - Soil moduli and damping factors for dynamic response analyses,Earthquake Engineering
Research Center. Report No. EERC 70-10. 1970.
[19] Stokoe II KH, Hwang SK, Lee JNK, Andrus RD - Effects of various parameters on the stiffness and damping
of soils at small to medium strains. In: Proceedings, International Symposium on Prefailure Deformation
Characteristics of Geomaterials, vol. 2, Japanese Society of Soil Mechanics and Foundation Engineering,
Sapporo, Japan, September 1994. p 785–816.
[20] Darendeli MB - Development of a new family of normalized modulus reduction and material damping curves,
Ph.D. Dissertation, The University of Texas at Austin, 2001.
[21] Amir-Faryar B, Aggour MS, McCuen RH - Universal model forms for predicting the shear modulus and
material damping of soils, Geomechanics and Geoengineering, 12(1), 2016. p 60–71.
https://doi.org/10.1080/17486025.2016.1162332
[22] Vinale F, d'Onofrio A, Mancuso C, Santucci de Magistris F, Tatsuoka F - The pre-failure behaviour of soils
as construction materials. In: Proceedings of the Second International Symposium on Pre-Failure
Deformation Characteristics of Geomaterials: Torino 99: Torino, Italy 28-30 September 1999, Volume 2.
[23] Løvholt F, Madshus C, Andersen KH - Intrinsic Soil Damping from Cyclic Laboratory Tests with Average
Strain Development. Geotech. Testing Journal, 43(1), 20170411. 2019. https://doi.org/10.1520/gtj20170411
 [24] Johansson J, Løvholt F, Madshus C - Procedures for estimating hysteretic foundation damping. In Frontiers
in Offshore Geotechnics III, 2015. p 1061–1066. CRC Press. https://doi.org/10.1201/b18442-157
[25] Kaynia A, Andersen K - Development of nonlinear foundation springs for dynamic analysis of platforms. In
Frontiers in Offshore Geotechnics III, 2015. p 1067–1072. CRC Press. https://doi.org/10.1201/b18442-158
[26] Page AM, Skau KS, Jostad HP, Eiksund GR - A New Foundation Model for Integrated Analyses of Monopile-
based Offshore Wind Turbines Energy Procedia, Elsevier BV, 2017, 137, 100-107.
[27] Hogeveen M, Hofstede H, Kaynia AM - Enhanced Kinematic Hardening Model for Load Dependent Stiffness
and Damping of Jack-up Foundations. In Proceedings: ASME 2018 37th International Conference on Ocean,
Offshore and Arctic Engineering. OMAE2018, Madrid, June 2018. Volume 9: Offshore Geotechnics;
Honoring Symposium for Professor Bernard Molin on Marine and Offshore Hydrodynamics, Paper No.
OMAE2018‑77285
[28] Oggiano L, Pierella F, Johansson J, Page AM - Comparison of Different Soil and Hydrodynamic Force
Models on a 13.2MW Offshore Rotor. ISOPE-I-19-172, 2019. Honolulu, Hawaii, USA: International Society
of Offshore and Polar Engineers. Retrieved from https://doi.org/
[29] De Alba P, Seed HB, and Chan CK - Sand liquefaction in large scale simple shear tests. J. Geotechnical Eng.
Div. ASCE 1976. 102:909-927.
[30] Boulanger RW, Idriss IM - CPT-based liquefaction triggering procedure. J Geotech and Geoenviron Eng
ASCE 2016;142(2). https://doi.org/10.1061/(ASCE)GT.1943-5606.0001388.
[31] Lu J, Elgamal A, Yan L, Law KH, Conte JP - Large-scale numerical modeling in geotechnical earthquake
engineering. International Journal of Geomechanics, ASCE 2011;11(6):490-503.
[32] Dafalias YF, Manzari MT - Simple plasticity sand model accounting for fabric change effects. Journal of
Engineering Mechanics 2004;130(6):622-634.
[33] Ziotopoulou K, Boulanger RW - Plasticity modelling of liquefaction effects under sloping ground and
irregular cyclic loading conditions. Soil Dynamics and Earthquake Engineering 2016; 84:269-283.
[34] Gobbi S, Lopez-Caballero F, Forcellini D - Numerical analysis of soil liquefaction induced failure of
embankments. In: Proceedings of 6th ECCOMAS Thematic Conference on Computational Methods in
Structural Dynamics and Earthquake Engineering. Rhodes Island, Greece; 2017.
[35] Rapti I, Lopez-Caballero F, Modaressi-Farahmand-Razavi A, Foucault A, Voldoire F. Liquefaction analysis
and damage evaluation of embankment-type structures. Acta Geotech 2018;13(5):1041-1059.
https://doi.org/10.1007/s11440-018-0631-z.
[36] Andrianopoulos KI, Papadimitriou AG, Bouckovalas, GD - Bounding surface plasticity model for the seismic
liquefaction analysis of geostructures. Soil Dyn. Earthquake Eng., 30(10), 895– 911.
[37] Andersen KH - Behaviour of clay subjected to undrained cyclic loading. Proc. 1st Int. Conf. on the Behaviour
of Offshore Structures, BOSS'76, Trondheim 1, 1976. p 392-403.
[38] Jostad HP, Grimstad G, Andersen KH, Saue M, Shin Y, You D - A FE Procedure for Foundation Design of
Offshore Structures – Applied to Study a Potential OWT Monopile Foundation in the Korean Western Sea.
Geotechnical Engineering Journal of the SEAGS & AGSSEA, 2014, Vol. 45, No. 4.
[39] Jostad HP, Grimstad G, Andersen KH, Sivasithamparam N - A FE procedure for calculation of cyclic
behaviour of offshore foundations under partly drained conditions. In: V. Meyer (Ed.), Frontiers in Offshore
Geotechnics III (ISFOG), Oslo, Norway, June 2015. Boca Raton: CRC Press. p 153-172.
[40] Lunne T, Andersen KH - Soft clay shear strength parameters for deepwater geotechnical design. In:
Proceedings of the International Offshore Site Investigation and Geotechnics Conference, 6. London 2007. p
151-176.
[41] Kjærnsli B, Valstad T, Høeg K - Rockfill dams: Design and Construction, Norwegian Institute of Technology,
ISBN 82-75598-012-11, Tapir. 1992.
[42] Duncan JM - Friction angles for sand, gravel, and rockfill. Kenneth Lee Memorial Seminar, Long Beach, CA,
2004.

View publication stats