Marine Structures 75 (2021) 102866

Contents lists available at ScienceDirect

Marine Structures
journal homepage: http://www.elsevier.com/locate/marstruc

A method to predict the torsional resistance of suction caisson

with anti-rotation fins in clay
Dengfeng Fu a, Zefeng Zhou b, c, *, Yue Yan a, Dhruba Lal Pradhan b, Jørn Hennig b
a
State Key Laboratory of Hydraulic Engineering Simulation and Safety, Tianjin University, Tianjin, 300072, China
b
Z1, Veritasveien 1, DNV GL, Oil & Gas, Z1, 1363, Høvik, Norway
c
Tuborg Parkvej 8, DNV GL, Renewable Energy, 2900, Hellerup, Denmark

A R T I C L E I N F O A B S T R A C T

Keywords: Suction caisson foundations have been increasingly employed as a primary solution to support the
Suction caisson offshore fixed- or mobile-structures. Due to harsh environment and complex force transferring of
Anti-rotation fin offshore structures, they are still being developed as to satisfy increasing requirements in strict
Soil flow mechanism
working scenarios. One of emerging challenges is a torsion-governing failure, which has been
Torsional resistance
Clay
observed in the oil and gas industry (i.e. significant multiple-inline-force-induced torsion) and in
the renewable energy field (i.e. non-coplanar tensile force induced torsion). This paper introduces
a novel suction caisson foundation, with anti-rotational fins assembled on the outer skin of
caisson. By a comprehensive numerical study, the evolution from local to global failure as the fin
numbers from single to multiple, is examined in the clayey soil deposit with effects of soil strength
heterogeneity, fins dimension, installation process and foundation-soil interface considered.
Based on these, a set of methods to estimate the ultimate torsional capacity of such novel caisson
is proposed, which starts from the gain in capacity for a single fin, and evaluate the changes of
gains in capacity as fin numbers, then identify the optimised anti-rotational capacity. Finally,
three key parameters (i.e. the required fin numbers, the available anti-rotational capacity and the
optimised anti-rotational capacity) with some critical considerations, recommendations and im­
plications have been concluded for design practice.

1. Introduction

Suction caisson foundations have been maturely and widely applied in the offshore oil and gas industry with remarkable com­
petences in installation cost (i.e. rapid and quiet installation) and bearing performance (i.e. additional tensile resistance from soil plug
under undrained loading condition). Recent applications of the suction caisson can be found in the offshore renewable industry, with
the diversity in the foundation forms, e.g. the tripod or quadruped suction caisson foundation systems reported by Hendrik 2017 and
[1]; the large-diameter bucket foundations with internal suction subdivisions [2–4] and the suction anchors group with multiple
mooring lines to hold the floating structure in place [5–7].
Suction caisson foundations are being increasingly favoured offshore and explicitly considered as an efficient and versatile offshore
foundation solution, which attracts more engineers and researchers to investigate and analyse the working mechanism through their

* Corresponding author. Z1, Veritasveien 1, DNV GL, Oil & Gas, Z1, 1363, Høvik, Norway.
E-mail addresses: dengfeng_fu@tju.edu.cn (D. Fu), zefeng.zhou@dnvgl.com (Z. Zhou), yue.yan@tju.edu.cn (Y. Yan), dhruba.lal.pradhan@dnvgl.
com (D.L. Pradhan), jorn.henning@dnvgl.com (J. Hennig).

https://doi.org/10.1016/j.marstruc.2020.102866
Received 25 July 2020; Accepted 8 September 2020
Available online 6 October 2020
0951-8339/© 2020 Published by Elsevier Ltd.
D. Fu et al. Marine Structures 75 (2021) 102866

Fig. 1. Typical torsion-related problems in offshore engineering.

2
D. Fu et al. Marine Structures 75 (2021) 102866

whole life, as to develop a set of comprehensive specifications and recommendations of guiding design, fabrication, installation, etc.
The estimation of geotechnical capacities for suction caisson foundation under ultimate limit states can be one of great significance for
application in practice. The classical bearing capacity theory [8–10] is still in use in design practice, where the multiplicative factors
are considered [11,12]. More recently, the failure envelope method has been recommended as an alternative use to represent the limit
states in three-dimensional loading space, such as API [13] and ISO [14]. It is capable of identifying the maximum allowable combined
capacities that the foundation can withstand, with derived expressions strongly depending on the foundation geometry, soil strength
profile, interface property, etc. A large body of studies focus on these using the upper or lower bound plasticity analyses (e.g.
Ref. [15–19]), finite element analyses (e.g. Ref. [20–30]), or physical modelling (e.g. Ref. [6,31–34]), with a comprehensive review of
these contributions provided by Randolph et al. [35] and Randolph et al. [36].
Although the previous studies have offered the significant basis for the design practice of suction caisson foundations in the course
of ultimate limit states, very little literature is available concerning the effect of non-coplanar loading on the capacity, especially in
cases where torsional loading could govern the failure of foundation. In fact, the torsion-governing failure has not been a rare case but
already observed and increasingly attracted wide attentions in the extensive application of suction caisson. Fig. 1 demonstrates three
typical loading cases for the suction caisson employed as an independent mooring anchor (Fig. 1a) or an anchor sharing mooring
system (Fig. 1b) to hold floating structures in place and as a shallow foundation to support the subsea modules (e.g. subsea distribution
unit and subsea manifold) (Fig. 1c). The tensile load along the mooring line (i.e. with a possible rotational angle ranging within ±5◦ in
general) or the multiple inline forces along each flowline connecting to the subsea module can generate the significant torsional
loading on the suction caisson foundation. It may dominate at failure rather than the conventional overturning moment or the coplanar
horizontal-moment combined loads. There is therefore an evident needed to optimise the geometry of suction caisson to increase the
anti-rotation capacity. It is a simply fact that the skin-soil shearing governs the failure of a suction caisson under the torsional loading,
so directly increasing diameter of caisson should be an option to enhance the anti-rotational capacity, yet the cost of fabrication and
installation may be significantly increased.
Alternatively, assembling additional fins along the outer skin of a caisson may be a more promising solution, when an equivalent
diameter is considered, Deq = D + 2s where D is the diameter of the suction caisson and s is the width of the fin, as illustrated in Fig. 2. If
the size of the caisson has already been designed to satisfy the requirement of the anti-rotation capacity (with its diameter indicated as
the red dash line), the proposed novel caisson (indicated as the yellow part) may potentially generate more torsional resistance.
Because the failure mode of such novel caisson, compared to a large-diameter conventional caisson, would be transferred from the
skin-soil to intact soil-soil shearing behaviour. And the reduction in fabrication material is significant. However, the build-up of such
global soil-soil shearing failure from local failure (see blue lines presented in Fig. 2) may be related to the number of fins, and the
magnitude of enhanced torsional resistance may be highly dependent on the width of fin, soil strength profile, soil-fin interface
property, etc.
As considered above, this novel suction caisson with anti-rotation fins is introduced in this paper. And the main objective is to
provide sufficient evidence for the feasibility of such caisson, and quantify the capability of increasing torsional resistance. This paper
initiates with the investigation of torsion-governing failure mechanism through the comprehensive finite element modelling analyses,
with considerations of fin numbers and dimension, soil property and interface properties. Then, a simple semi-analytical method is
proposed to assess and optimise the anti-rotation capacity with several critical considerations, recommendations and implications
provided for design practice.

2. Finite element modelling

The three-dimensional finite element modelling analyses are conducted using the commercial finite element software ABAQUS
[37]. This section provides a detailed description of the finite element modelling and an overview of the analyses performed.

Fig. 2. Concept of increasing the torsional resistance of suction caisson by additional anti-rotation fins.

3
D. Fu et al. Marine Structures 75 (2021) 102866

2.1. Foundation model

A typical suction caisson geometry with anti-rotation fins considered in this study is illustrated in Fig. 3. The skirted suction caisson
of diameter D = 6 m with the length L of 9 m, represents a typical dimension of suction caisson expected to be used offshore. The anti-
rotation fins are arranged evenly around the periphery of the foundation, with the fin numbers varying between one and eight. The

Fig. 3. Meshing technology in finite element model.

4
D. Fu et al. Marine Structures 75 (2021) 102866

length of fins L0 is assumed to be the same as the foundation penetration depth, unless prescribed otherwise. The width of the fins to the
foundation diameter ratio s/D varies between 0.01 and 0.5 (i.e. 0.01, 0.03, 0.05, 0.07, 0.09, 0.1, 0.2, 0.25 and 0.5). The foundation
skirt and fin thickness are both set to 0.03 m, corresponding to a thickness to diameter ratio t/D of 0.005 for the foundation model. The
foundation shown in Fig. 3 represents a suction caisson with three fins, with s/D of 0.5 and L0/L of 1.
The foundation is modelled as a rigid body, with the unit weight set equal to the soil for the ease of equilibrium under the geostatic
loading. Interface between the skirt tip (skirt internal and external vertical sides) and soil is fully bonded – that is, rough in shear with
no detachment permitted. More considerations for the interface properties are described later.
The position of the reference point is selected at the center of the suction caisson foundation at the skirt tip level, although it will not
influence the torsional resistance discussed in this paper.

2.2. Soil model

The soil domain, which is shown in Fig. 3, extends 7(s + D) from the radius direction and 3L beneath the foundation tip to avoid
boundary effects. The soil base is fixed and the lateral boundary only allows vertical displacements. The minimum thickness of the
elements is 0.002D, which are used adjacent to the external wall of the suction caisson and anti-rotation fins. The typical mesh shown
in Fig. 3 comprises around 80,000 for the example model. First order, fully integration, hybrid hexahedral elements (ref. C3D8H,
ABAQUS element library) are used in the analyses.
The undrained material response of the soil was represented as a linear elastic and perfectly-plastic material, obeying the Tresca
criterion. The elastic properties were set as undrained Young’s modulus, E = 500su, and Poisson ratio, ν = 0.49. Two limiting soil
strength profiles are considered, which represents a normalised sum/kD of 0 (normally consolidated soil) and infinite (homogeneous
soil). The effective unit weight of the soil γ′ is set as 6 kN/m3.

Fig. 4. Validation of finite element modelling method (L/D = 1.5).

5
D. Fu et al. Marine Structures 75 (2021) 102866

2.3. Loading method and parametric study

Geostatic equilibrium is setup following the application of self-weight stresses. Then the torsional displacement-controlled tests are
undertaken to obtain the torsional resistances. Most numerical analyses are undertaken with the foundation-soil interface configu­
ration already described in section 2.1.
However, to examine torsional resistance of the novel caisson, three groups of analyses are undertaken with the following setting:

(1) Effect of installation process: during the penetration process, the soil surrounding the suction caisson is remoulded, and the
localised remoulded soil is normally in the thickness of t (as recommended by Ref. [38–40]). For modelling such effect on the

Fig. 5. Load-displacement response for a typical suction caisson with anti-rotation fins (s/D = 0.25, L/D = 1.5).

6
D. Fu et al. Marine Structures 75 (2021) 102866

torsional capacity, the soil strength of su/St (St is the soil sensitivity, and taken as 5 in this study for a typical value of marine
deposits ([41,42]) is specified to a thin layer of soil around the skin wall of the finned suction caisson foundation, as shown in
Fig. 3.
(2) Effect of tension gap and interface roughness: the contact properties of the skirt-soil, and fin external vertical sides-soil are taken
as frictionless with separation allowed (zero tension condition). The choice is made as to obtain conservative values of torsional
capacity of the foundation.
(3) Effect of fin length: the assembled fin length (L0) is changed to one third and two thirds of the caisson full length, with the fin
welded from the caisson tip level.

2.4. Validation of the finite element model

No solution for the torsional resistance of a suction caisson with fins is publicly available to date. Therefore, the numerical model
was validated against an analytical solution for the torsional resistance of the conventional suction caisson foundation. As known, the
failure of the torsional shearing occurs at a thin soil shear band beneath the foundation base and around external skirt wall. The
theoretical torsional resistance therefore can be expressed as follows:
for homogeneous soil:

αsu π D2 L su πD3
Thomo = + (1)
2 12

for normally consolidated soil:

αkπD2 L2 kLπD3
TNCsoil = + (2)
4 12

where α is the partially-remoulded factor of soil strength, representing the effect of installation process on the torsional capacity, with
α = 1 and 1/St for intact and fully remoulded soil conditions recommended by DNV GL [43]; su is the shear strength for the homo­
geneous case; k is the undrained shear strength gradient with depth for normally consolidated soil cases.
Fig. 4 shows the calculated value of the normalised capacity factors from the finite element analyses and the analytical solutions for
a suction caisson with L/D = 1.5. Note, the results are presented in terms of non-dimensional load capacity factor T/ADsu,tip, where A is
the plane area (= πD2/4), su,tip is the undrained shear strength at the tip level (= su for the homogeneous soil; and = kL for the normally
consolidated soil). Comparison of the results show close agreement, with the maximum difference within 1.5%. These results validate
the finite element meshing technology, which is then used to construct the anti-rotation fin-soil interfaces of the novel foundation.

2.5. FEM modelling results

Fig. 5 reports responses of torsional loading against rotation angle in both non-dimensional (T/ADsu,tip) and normalised (T/T0,
where T0 is the ultimate torsional capacity, indicated by the plateau of torque-displacement curves) groups. Cases of both homogenous
and normally consolidated soil with s/D of 0.25 and fin number, N, varying from 0 to 8 are considered. Note that the fully-bonded
constraint between foundation and soil is considered in this section.
As evident from non-dimensional results, considerable increase in the torsional resistance can be obtained by increasing fin
numbers, with the maximum enhancement approximately of 2.4 times the non-finned results. From normalised results, the build-up of
torque for the cases with different fin numbers shows the interesting tendency: (1) less fin numbers give a softer mobilisation and, (2)
the increasing fin numbers enable the mobilisation to be stiffer and finally approach the non-finned results. To interpret it, the evo­
lution of soil flow mechanism with increasing fin numbers is demonstrated in Fig. 6. The significant change in deformation contours

Fig. 6. Typical displacement contours from top view for a typical suction caisson with anti-rotation fins (s/D = 0.25, L/D = 1.5) in homoge­
neous soil.

7
D. Fu et al. Marine Structures 75 (2021) 102866

can be observed and classified into three typical patterns: (1) one-fin triggers the local failure mode, with the active and passive earth
pressure dominating at failure; (2) two-fins modifies the local failure mode, with more surrounding soil being involved; (3) more fins
(from N = 3) specify the shearing zone and form the global failure, due to mutual constraint effect of fins. Note that the base shear
mechanism is not of concern, as its contribution is relatively small compared to periphery shearing (L/D > 1). These manifest that more
soil can be mobilised around the fin, leading to an increase in torsional resistance, but the local failure (i.e. active and passive earth
pressure involved) can significantly reduce the torsional stiffness. These reduction effect can be negligible until the sufficient fin
numbers are taken as to perform the global failure.
Fig. 7 shows the enhancement of torsional resistance for the suction caisson with fins against different fin width ratio s/D compared
to the non-finned foundation. Both homogeneous and normally consolidated soil conditions are included in the normalised form of T/
T0. For all cases, a tiny increase response can be found when the fin number is larger than 4, implying a global shearing failure
(described above) is progressively approached with the further increase in fin numbers. Larger s/D can give higher enhancement in T/
T0, with the same fin number considered, yet larger s/D seems to require more fins to develop global failure mechanism. The maximum
enhancement in torsional resistance is approximately 4.32 and 4.64 for homogeneous and normally consolidated soil, respectively,
which are observed in the cases of s/D = 0.5 with fin number of 8.

3. Critical consideration

The focus of this section is on several critical considerations for the application of the finned suction caisson foundation. The
reduction in the soil strength due to the installation process, the foundation-soil interface property (roughness and separation status),
and the assembled length ratio of fin (L0/L, where L0 is the desired fin length) are separately discussed in the following three
subsections.

Fig. 7. Torsional resistance of suction caissons with anti-rotation fins.

8
D. Fu et al. Marine Structures 75 (2021) 102866

3.1. Effect of installation process

As presented by the previous study [38,44], the soil surrounding the caisson can be remoulded after installation, leading to the
remoulded soil strength close to su/St and the typical remoulded zone within a wall-thickness width t. When additional fins are added
on the caisson, the soil surrounding the fins could be remoulded as well. To investigate these effects on the anti-rotation capacity, this
study takes a 30 mm-thin (equal to the thickness of skirt) soil layer to cover the vertical outer skin of both caisson periphery and fins,
with undrained shear strength specified as 1/St of the initial strength. This is consistent with strain-softening soil strength due to effect
of the installation recommended by DNV GL [43]; and such consideration should be practical and conservative in design.
A suit of simulations set up following above are implemented with s/D = 0.25 for homogeneous soil condition. Fig. 8a shows a
series of normalised displacement contour diagrams from top view. It is clear that the mobilised soil region induced by a single fin
significantly shrinks, and even for a two-fins case, two separate flow zones stand aside without any combination. Until three fins are
assembled, the flow zones merge as a global mode. These are significantly different from that observed in Fig. 6 for the cases without
the installation effect. After that, the soil deformation patterns for each fin number are similar to the cases with and without the
installation effect. Due to changes in the soil failure mechanism, especially at N < 3, the enhancement in anti-rotation capacity can be
influenced by the reduction in the intact soil strength surrounding the fins. The corresponding results of T/T0 are reported in Fig. 8b,
where the cases with and without installation effect are included for comparison. As predicted, the observable decrease in T/T0 can be
found for N < 3. But almost identical responses are obtained when N ≥ 3, implying the installation effect can be minimum once the
global failure mode is mobilised, since the soil-soil shearing mode will elude the already remoulded soil regions (i.e. soil surrounding
the caisson outer periphery, and front and rear side of a fin).

3.2. Effect of tension gap and interface roughness

The separation phenomenon is often observed for offshore foundations, for instance, Randolph et al. [45] reported an immediate
gap formed at the rear side for a laterally-loaded suction anchor in the lightly over-consolidated soil during centrifuge tests. The soil on
the active side might not be subjected to the tensile load, resulting a separation with foundation, i.e. tension gap. It could lead to a
reduction in foundation capacity. For the rotation behaviour of fins, the tension gap may occur as well. Besides, the interaction

Fig. 8. Effect of installation process on the torsional resistance for suction caissons with anti-rotation fins (s/D = 0.25) in homogeneous soil.

9
D. Fu et al. Marine Structures 75 (2021) 102866

between soil and fins during rotation could be influenced by the interface roughness.
To investigate these effects on the anti-rotation capacity, insights into soil flow mechanisms with different foundation-soil interface
properties are firstly examined, where an anti-clockwise rotation is applied on the foundation. Only homogeneous soil condition is
considered here, since no tension gap exists in the normally consolidated soil, which is consistent with the previous observation [45].

Fig. 9. Effect of interface property on the torsional resistance for suction caissons with anti-rotation fins (s/D = 0.25) in homogeneous soil.

10
D. Fu et al. Marine Structures 75 (2021) 102866

Fig. 9. (continued).

The fully rough interface case without considering tension gap can refer to Fig. 6, since they give almost identical responses, and can be
taken as a reference representing the peak torsional resistance. When the effect of the tension gap is involved with the foundation-soil
interface specified as fully smooth, as presented in Fig. 9a, progressive changes in soil failure mechanism from local mode (i.e. N < 5) to
global mode (i.e. N ≥ 5), can be explicitly observed. Several isolated flow zones are localised on one side of fin when N < 5, where the
tension gap effect is significant, with no soil behind the fin mobilised. Another set of cases, representing the most critical field condition
(i.e. a thin layer of fully-remoulded soil (α = 1/St) in contact with the fully smooth interface), are demonstrated in Fig. 9b. As evident,
they indicate the most conservative responses. The required fin number of 5 is observed to mobilise a global failure mode, which is
higher than the fin number of 3 to achieve a global failure mode for the fully rough interface case, and similar to the case of fully
smooth interface with tension gap.
The corresponding results of T/T0 are presented in Fig. 9c with s/D of 0.25 in homogeneous soil. The fully rough interface with no
tension gap gives the largest torsional resistance, while combining the effects of tension gap, fully smooth interface and soil remoulding
gives the lowest torsional resistance. The response of fully smooth interface with tension gap falls in between the above two. They all
perform a close global failure from the fin number around 5. Although the combing effects lead to a slightly slow evolution towards the
global failure mode, the discrepancy of torsional resistances is only with 5% from N of 4.

3.3. Effect of the assembled fin length L0

With consideration from the geotechnical capacity design aspect, a full-length equal to the height of the caisson used to size the fin
dimension could be the most effective method. Yet, the upper part of fins close to the caisson lid might trigger scouring at the shallow
depth [46,47]. A proper shorten length of fin should be therefore considered and only assembled from the bottom of the caisson.
However, the reduction in the fins could result in a lower enhancement in torsional resistance. Thus, this section examines the per­
formance of fins with reduced length on the anti-rotation capacity.
Fig. 10a demonstrate two groups of failure mechanisms under the assembled length ratios, L0/L of 0.33 and 0.67 respectively,
where s/D of 0.25 and homogeneous soil conditions are considered. To identify the difference from the mechanism of assembling the
full-length fins (each fin paths through the caisson height, i.e. L0/L of 1), the soil flow patterns from one side-view and two (horizontal)
cut views are presented. From the side view, regardless of the fin numbers and the assembled fin lengths, the skin-soil shearing mode in
the vertical plane dominates at the top of fins, and the soil-soil shearing mode still prevails and crosses the entire assembled fin. From
two cut views, the top surface of fin generates a new annular top shear, which can be thought as the soil-soil shearing mode. It is almost
identical to the base shear, except for the case of fin number of 1 and L0/L of 0.67. For each L0/L, the global failure mode forms until fin
number reaches 4, localised at the region of assembled fins. As described above, the main differences of failure mechanism between
full-length and assembled-length fins are: (1) the extent of global failure mode indeed shrinks due to the reduction in the fin length; (2)
an extra annular top shear is generated by the top surface of assembled fins.
Their resultant responses of anti-rotation capacity are reported Fig. 10b. For each L0/L, the torsional resistance gives a similar
tendency with a tiny increase when N ≥ 4, which implies the evolution of failure mechanism towards the global failure with fin
numbers is similar. For the same fin numbers, considerable reduction in torsional resistance against L0/L is observed, indicating the

11
D. Fu et al. Marine Structures 75 (2021) 102866

Fig. 10. Effect of fin length on the torsional resistance for suction caissons with anti-rotation fins (s/D = 0.25).

12
D. Fu et al. Marine Structures 75 (2021) 102866

Fig. 10. (continued).

compensation effect induced by the annular top shear is not sufficient.

Additionally, both failure mechanisms and their related torsional resistance in normally consolidated soil are also shown in Fig. 10c
and d. Compared to the homogeneous soil condition, similar observations in the evolution of failure mechanisms against fin numbers
can be found. However, the responses of torsional resistance give an interesting tendency: almost identical responses are captured for
L0/L of 0.67 and 1. This implies the annular top shear failure mode is pronounced and almost compensates the reduction induced by the
decrease in the fin length.

4. A method to predict anti-rotation capacity

This section describes a method to predict the anti-rotation capacity by adding the fins on the suction caisson. The method is
proposed with the aim to solve two things of concern in the design practice: (1) to evaluate the minimum fin number ensuring the
global failure mechanism to be mobilised; (2) to estimate the corresponding anti-rotation capacity. The method includes three main
components: the enhancement in anti-rotation capacity due to a single fin; the evolution of enhancement against fin numbers and the
available anti-rotation capacity.
Firstly, to evaluate the contribution of single fin to the torsional resistance ΔTN=1, the behaviour of rotating the single fin around an
axis of symmetry is thought to be similar to the translation of the laterally-loaded deep foundation at the small strain level, as
demonstrated in Fig. 11. Thus, ΔTN=1 can be expressed as
∫
s+D
ΔTN=1 = δN=1 Np,N=1 su s dz (3)
2

where δN=1 is a scaling factor, as discussed later; and Np,N=1 in Eq. (3) can be derived following Zhang et al. [3,4] and Fu et al. [48,49]:
Np,N=1 = 2Np0,N=1 ≤ Npd (4)

where,
[ ( )0.6 ]1.35
z
Np0,N=1 = N1 − (N1 − N2 ) 1 − ≤ Npd
H(s + 0.5D)

N1 = 11.94

N2 = 3.22
(s )
um
H = 16.8 − 2.3log ≥ 14.5
kD

13
D. Fu et al. Marine Structures 75 (2021) 102866

Fig. 10. (continued).

14
D. Fu et al. Marine Structures 75 (2021) 102866

Fig. 10. (continued).

Fig. 11. Schematic plots for scaling the behaviour from translation of laterally loaded deep foundation to rotating of the single fin at small
strain level.

Npd = 11.94

It is notable that both active and passive earth pressures are included in Eq. (4), since the global failure mode is indeed expected in

15
D. Fu et al. Marine Structures 75 (2021) 102866

design, where the effect of tension gap can be neglected. The coefficient of N1 and N2 represent the typical capacity factors at the deep
and shallow depths of soil, and H stands for the normalised depth at which the shallow and deep mechanism shift entirely. These are
captured from numerical analyses [3,4].
As acknowledged that, the failure mode might be modified when the fin is connected to the caisson. The main difference is the
change in the mobilised soil region. The reduction in extent of the mobilised soil at failure should be considered, since the suction
caisson hinders the sideways soil mobilisation of the fin. Thus, a scaling factor δN=1 is introduced here (see equation (3)). Least squares
regression analyses are performed over the entire database (both homogeneous and normally consolidated soil), yielding the ex­
pressions of δN=1 with s/D:
s
δN=1 = 0.166log +1 (5)
D
Note δN=1 also covers the transition of plane strain condition, since Eq. (4) is derived from the scenario of a laterally-loaded pile.
The estimation of δN=1 using Eq. (5) is compared to the results processed from numerical analyses, as shown in Fig. 12a. Accordingly,
the results of ΔTN=1/T0 calculated by Eqs. (1)–(5) are reported in Fig. 12b for a comparison with numerical results. As evident, they
generally give a good agreement.
Secondly, to estimate the contribution of multiple fins, the evolution of enhancement against fin numbers is derived by introducing
a build-up coefficient ζN.
ΔTN = ζN ΔTN=1 (6)
ζN simply bridges the accumulative enhancement due to multiple fins ΔTN and the enhancement from single fin ΔTN=1. It reveals
the changes in mechanism from single to multiple fins, where the increase in fins leads to a mutual constraint phenomenon, thereby a
decreasing enhancement compared to that generated by the single fin. Here, all numerical results for the normalised enhancement

Fig. 12. The contribution of single fin on the anti-rotation capacity.

16
D. Fu et al. Marine Structures 75 (2021) 102866

Fig. 13. The evolution of enhancement (ΔTN+1 – ΔTN)/ΔTN=1 against fin numbers.

(ΔTN+1 – ΔTN)/ΔTN=1 (N starts from 1) due to fins for each s/D are extrapolated in Fig. 13. It is clear that a build-up coefficient ζN
against fin numbers exist, and give a reasonable prediction for all results, expressed as.
For homogeneous soil:
( )2.3
∑
i=N i− 1

(7)
1.15
ζN = 0.5
i=1

For normally consolidated soil:

( )2.1
∑
i=N i− 1

(8)
1.4
ζN = 0.5
i=1

Therefore, the anti-rotation capacity of the caisson foundation with fins TN can be simply expressed by the superposition of the
torsional resistance of non-finned caisson T0 and the enhancement in resistance due to fins, ΔTN:
TN = T0 + ΔTN ≤ Tava (9)

where Tava is the available anti-rotation capacity for suction caisson with fins, associated with a global failure mechanism, and
expressed as.
for homogeneous soil

su π(D + 2s)2 L su π (D + 2s)3

Tava = + (10)
2 12

for normally consolidated soil

17
D. Fu et al. Marine Structures 75 (2021) 102866

kπ(D + 2s)2 L2 kLπ(D + 2s)3

Tava = + (11)
4 12
As known, for an identical s/D, the anti-rotation capacity TN will infinitely approach Tava with increasing fin numbers. Thus, it is
convenient for design practice that a coefficient of 0.95 can be applied to Tava (here 0.95Tava can be taken as the design load Treq), as to
determine the required fin numbers.
Fig. 14 summarises the normalised anti-rotation capacity T/T0 of the suction caisson with fins from both the numerical analyses and
the predicted by Eq. (9) in homogeneous and normally consolidated soil. It is clear that the proposed method was shown to simulate
well the torsional capacity of the novel suction caisson, with maximum estimation discrepancies of within 2% and 1% for homoge­
neous and normally consolidated soil, respectively.

5. Example application

This section demonstrates a practice example, where the considerable torsional loading on the suction caisson foundation governs
the design. How the anti-rotation fins work to increase the torsional resistance and the relevant calculation procedure implemented to
optimise the torsional capacity are outlined in details.
The basic information necessary for design can be described as: a steel-made suction caisson foundation with diameter of 6 m,
thickness of 30 mm, and the length of 9 m is founded in a typical clayey soil deposit. The detailed undrained shear strength profiles (su
against soil depth z) are interpreted from CPT tests, and shown in Fig. 15a. Here the total stress method is used with the parameter Nkt
of 15 and 20 for the high and low estimation of su.
Using the low estimate of su, the ultimate torsional resistance of the non-finned foundation is 3.6 MNm according Eq. (2). Because
the designed torsional loading of 5.4 MNm is applied on the foundation, it is evident the increasing the foundation geometry is
necessary. Here, two measures are implemented: (i) only the diameter of the suction caisson foundation is increased; and (ii) sizing the
additional fins to the original foundation.

Fig. 14. Finite element results and prediction of torsional bearing capacities for the suction caisson with fins.

18
D. Fu et al. Marine Structures 75 (2021) 102866
Fig. 15. Example application.
19
D. Fu et al. Marine Structures 75 (2021) 102866

For the former, the required diameter Dreq is 7.08 m according to Eq. (2). For the latter, the required width s of fin can be
determined by the available torsional resistance Tava (Tava = Treq/0.95), which leads to s = 0.6 m. Then the required numbers of fin can
be examined by calculating the anti-rotation finned caisson using Eq. (8), with the detailed procedure illustrated in Fig. 15b. As
evident, Nreq = 3, yielding a TN = 5.7 MNm.
Note, although increasing the fin width and fin numbers can generate more gains in the anti-rotation capacity, excessive increase in
the dimension of fin and fin numbers may lead the installation significantly costly and even unreachable. Thus, the installation process
for both non-finned and finned suction caisson foundations need to be examined. According to DNV GL method [43] for installation of
suction caisson, both the required and available suction pressure are calculated and shown in Fig. 15c. As evident, compared to the
results for the non-finned foundation (red solid line), the finned foundation with s/D = 0.1 (red dash line) only gives a slight increase in
the required suction pressure. The other three cases with s/D = 0.05, 0.25 and 0.5 are also included for comparison. Although the early
penetration depths by the self-weight of foundations are similar, the disparities of required suctions with the increasing s/D are
observable. e.g. for the target installation depth of 9 m, a suction pressure of around 44.9 kPa is required for s/D = 0.5while a suction of
33.1 kPa is needed for non-finned foundation. So, the excessive increase in fin width may result in a considerable rise in the required
suction pressure, as well the initial self-weight penetration depth (though is insusceptible of s/D in this specified site condition).
However, such rise in the required suction is completely allowable with sufficient safety margin away from the critical available
suction pressure (the solid black line, indicating a maximum suction associated with the internal soil plug failure mode).
A further examination for the efficiency of assembled fins is made by the proposed method, and the results are denoted as the
efficiency ratio, which is the anti-rotational capacity divided by the weight of foundations T/W′ (W′ is submerged weight of foun­
dation, here the submerged unit weight of steel γ′ steel is taken as 68.7 kN/m3). Four different diameters of foundation (D = 5, 6, 7 and 8
m) with s/D = 0 to 0.5 and N = 3 are considered, and other relevant parameters for foundation and soil properties necessary for
calculations are same as those in the example described above. As illustrated in Fig. 15d, the non-finned foundations (s/D = 0) show the
lowest efficiency while the widest fin performs the highest efficiency. Thus, it is manifest that the method using additional fins to
enhance the torsional resistance can give a cost-effective solution.

6. Conclusion

This paper presents a novel suction caisson foundation conceptually, with its promising anti-rotational performance captured from
the additionally-assembled peripheral fins examined. From the comprehensive insights into the evolution of soil flow mechanisms
against fins numbers under both homogenous and normally consolidated soil conditions, the main achievements useful for the design
practice can be concluded as:

• The single fin yields the maximum enhancement in the anti-rotational capacity, yet such gain cannot be maintained as the fin
numbers increase, since the significant mutual constraint results in the progressive formation of global failure mode. It implies that
the required (effective) fin numbers should be determined in design as to ensure a costly solution in fins.
• The peripheral fins can be expected to trigger the global failure with mainly soil-soil shearing mode dominated, where the intact
soil strength can be mobilised without the installation effect on the anti-rotational capacity necessary to be considered, and the
effect of interface property on the capacity is minimum as well. It suggests an available anti-rotational capacity for the proposed
suction caisson foundation, which is useful to identify the width of fins.
• The reduction of the fin length may lead to a certain decrease in anti-rotational capacity, which strongly depends on the soil
strength property for the layer without fins. If is linearly-increasing strength profile (i.e. normally-consolidated soil), a possibly
equivalent (even larger) anti-rotational capacity can be obtained, compared with full and partial length finned caisson. This rec­
ommends a certain length of fin from the top of caisson is possible to be cut off, ensuring a probably better capacity and anti-scour
capability.

Based on these examinations of the proposed finned suction caisson foundation, an analytical method to predict its optimised anti-
rotational capacity is provided as to facilitate the sizing design in practice.
It is worthy to be mentioned here, the coupling effect of different loading components indeed exists due to the harsh environmental
and complex structural forces transferring. This study focuses on the ultimate anti-rotational capacity without any other directional
loading considered, but it provides the fundamental investigations of enhancing anti-rotational capacity for this novel foundation, and
significant basis and valuable recommendations for both research and industry communities.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Acknowledgements

The authors would like to acknowledge the support received from the National Natural Science Foundation of China (No.
41877214; No. 51709198; No. 51890913).

20
D. Fu et al. Marine Structures 75 (2021) 102866

Notation

A plan area of the suction caisson

D diameter of suction caisson
Dreq required diameter of an equivalent SCF
E Young’s modulus of soil
H normalised depth at which shallow and depth mechanism shift
k undrained shear strength gradient
L length of suction caisson
L0 length of fins
N number of fins
Np,N=1 bearing capacity factor of torsional loaded one fin around the suction caisson when suction is assumed
Np0,N=1 bearing capacity factor of torsional loaded of one fin around the suction caisson in idealised weightless soil
Npd limiting bearing capacity factor mobilised in a localised flow-around mechanism
N1 Typical capacity factor at deep depths of soil
N2 Typical capacity factor at shallow depths of soil
s fin width
su undrained shear strength
su,tip undrained shear strength at the skirt tip level
St soil sensitivity
t skirt thickness
T torsional resistance
Tava available anti-rotation capacity for suction caisson with fins associated with a global failure mechanism
Tult ultimate torsional resistance
T0 ultimate torsional resistance of the suction caisson
α partially-remoulded factor of soil strength
δN=1 scaling factor
ζN accumulative enhancement factor due to multiple fins
θ rational angle
ν Poisson ratio of soil

References

[1] Skau KS, Grimstad G, Page A, Eiksund G, Jostad HP. A macro-element for integrated time domain analyses representing bucket foundations for offshore wind
turbines. Mar Struct 2018;59:158–78.
[2] Ding H, Liu Y, Le C, Zhang P. Model tests on the bearing capacity of wide-shallow composite bucket foundations for offshore wind turbines in clay. Ocean
Engineering; 2015. 103:114~122.
[3] Zhang Y, Andersen KH, Tedesco G. Ultimate bearing capacity of laterally loaded piles in clay - some practical considerations. Mar Struct 2016;50(2016):260–75.
[4] Zhang P, Guo Y, Liu Y, Ding H. Experimental study on installation of hybrid bucket foundations for offshore wind turbines in silty clay. Ocean Eng 2016;114:
87–100.
[5] Gaudin C, O’Loughlin C, Duong MT, Herduin M, Fiumana N, Draper S, Wolgamot HA, Zhao L, Cassidy MJ. New anchoring paradigms for floating renewables. In:
Proceedings of the 12th European wave and tidal energy conference. United Kingdom: Technical Committee of the European Wave and Tidal Energy
Conference; 2018. p. 9.
[6] Herduin M, Gaudin C, Johanning L. Anchor sharing in sands: centrifuge modelling and soil element testing to characterise multi-directional loadings. In:
Offshore geotechnics; honoring symposium for professor bernard molin on marine and offshore hydrodynamics. vol. 9. The American Society of Mechanical
Engineers (ASME); 2018. 9 pp. V009T10A030.
[7] Hegseth JM, Bachynski EE, Martins J. Integrated design optimization of spar floating wind turbines. Mar Struct 2020;72:102771.
[8] Davis EH, Booker JR. The effect of increasing strength with depth on the bearing capacity of clays. Geotechnique 1973;23(4):551–63.
[9] Prandtl L. Eindringungsfestigkeit und festigkeit von schneiden. Angew Math U Mech 1921;1(15).
[10] Terzaghi K. Theoretical soil mechanics. New York: Wiley; 1943.
[11] Brinch Hansen J. A revised and extended formula for bearing capacityvol. 98. Conpenhagen: The Danish Geotechnical Institute; 1970. p. 5–11.
[12] Meyerhof GG. The bearing capacity of foundations under eccentric and inclined loads. In: Proceedings of the 3 rd ICSMFE, zurich, vol. 1; 1953. p. 440–5.
[13] API. Recommended practice 2GEO geotechnical and foundation design considerations. 1st end. Washington, DC, USA: American Petroleum Institute; 2011.
[14] ISO. ISO 19901-4: 2016 (en). Petroleum and natural gas industries specific requirements for Offshore Structures – Part 4: geotechnical and foundation design
considerations. Geneva, Switzerland: International Standards Organisation; 2016.
[15] Bransby MF, Randolph MF. Combined loading of skirted foundations. Geotechnique 1998;48(5):637–55.
[16] Bransby MF, Randolph MF. The effects of embedment on the undrained response of caisson foundations to combined loadings. Soils Found 1999;39(4):19–34.
[17] Houlsby GT, Puzrin AM. The bearing capacity of a strip footing on clay under combined loading. Proc Royal Soc Londn Ser A 1999;455:893–916.
[18] Randolph MF, Puzrin AM. Upper bound limit analysis of circular foundations on clay under general loading. Geotechnique 2003;53(9):785–96.
[19] Salgado R, Lyamin AV, Sloan SW, Yu HS. Two and three-dimensional bearing capacity of foundations in clay. Geotechnique 2004;54(5):297–306.
[20] Bransby MF, Yun GJ. The undrained capacity of skirted strip foundations under combined loading. Geotechnique 2009;59(2):115–25.
[21] Fu D, Gaudin C, Tian Y, Cassidy MJ, Bienen B. Uniaxial capacities of skirted circular foundations in clay. J Geotech Geoenviron Eng 2017;7(143).
[22] Fu D, Gaudin C, Bienen B, Tian Y, Cassidy MJ. Combined load capacity of a preloaded skirted circular foundation in clay. J Geotech Geoenviron Eng V 2018;144
(11).
[23] Gourvenec S. Failure envelopes for offshore shallow foundations under general loading. Geotechnique 2007;57(9):715–28.
[24] Gourvenec S. Effect of embedment on the undrained capacity of shallow foundations under general loading. Geotechnique 2008;58(3):177–85.

21
D. Fu et al. Marine Structures 75 (2021) 102866

[25] Gourvenec S, Randolph MF. Effect of strength non-homogeneity on the shape and failure envelopes for combined loading of strip and circular foundations on
clay. Geotechnique 2003;53(6):575–86.
[26] Taiebat HA, Carter JP. Numerical studies of the bearing capacity of shallow foundations on cohesive soil subjected to combined loading. Geotechnique 2000;50
(4):409–18.
[27] Taiebat HA, Carter JP. Bearing capacity of strip and circular foundations on undrained clay subjected to eccentric loads. Geotechnique 2002;52(1):61–4.
[28] Taiebat HA, Carter JP. A failure surface for circular footings on cohesive soils. Geotechnique 2010;60(4):265–73.
[29] Vulpe C, Gourvenec S, Power M. A generalised failure envelop for undrained capacity of circular shallow foundations under general loading. Géotech Lett 2014;
4(3):187–96.
[30] Vulpe C, Gourvenec S, Leman B, Fung KN. Failure envelope approach for consolidated undrained capacity of shallow foundations. J Geotech Geoenviron Eng
2016:1–13. 04016036.
[31] Bienen B, Byrne BW, Houlsby GT, Cassidy MJ. Investigating six-degree-of-freedom loading of shallow foundations on sand. Geotechnique 2006;56(6):367–79.
[32] Bienen B, Gaudin C, Cassidy MJ. Centrifuge tests of shallow footing behaviour on sand under combined vertical torsional loading. Int J Phys Model Geotech
2007;7(2):1–22.
[33] Byrne BW, Houlsby GT. Investigating 6 degree of freedom loading on shallow foundations. In: Proceedings of international symposium on frontiers in offshore
geotechnics (ISFOG), perth; 2005. p. 477–82.
[34] Cassidy MJ, Cheong J. The behaviour of circular footings on sand subjected to combined vertical-torsion loading. Int J Phys Model Geotech 2005;5(4):1–14.
[35] Randolph MF, Cassidy MJ, Gourvenec S, Erbrich C. Characterisation of soft sediments for offshore applications. In: Proceedings of the 2nd international
conference on site characterisation, porto, Portugal; 2005. p. 209–31.
[36] Randolph MF, Gaudin C, Gourvenec S, White DJ, Boylan N, Cassidy MJ. Recent advances in offshore geotechnics for deep water oil and gas developments.
Ocean Eng 2011;38(7):818–34 (2011).
[37] Dassault Systèmes. Abaqus analysis users’ manual. Providence, RI, USA: Simula Corp; 2010.
[38] Andersen KH, Jostad HP. Shear strength along outside wall of suction anchors in clay after installation. In: Proceedings of the twelfth international offshore and
polar engineering conference. Japan: Kitakyushu; 2002.
[39] Renzi R, Maggioni W, Smits F, Manes V. A centrifugal study on the behaviour of suction piles. In: Int. Conf. Centrifuge 91, boulder, Colorado. Balkema,
rotterdam, ISBN 90 6191 1931. p. 169–76.
[40] Sagaseta C, Whittle AJ, Santagata M. Deformation analyses of shallow penetration in clay. Int J Num and Anal. Meth. in Geomech. 1997;21:687–719.
[41] Zhou Z, O’Loughlin CD, White DJ, Stanier S. Improvements in plate anchor capacity due to cyclic and maintained loads combined with consolidation.
Geotechnique 2020;70(8):732–49.
[42] Zhou Z, White DJ, O’Loughlin CD. The changing strength of carbonate silt: parallel penetrometer and foundation tests with cyclic loading and reconsolidation
periods. Can Geotech J 2020. https://doi.org/10.1139/cgj-2019-0066 [ahead of print].
[43] DNVGL. RP E303, Geotechnical design and installation of suction anchors in clay. Oslo, Norway: DNVGL; 2017.
[44] Xiao Z, Fu D, Zhou Z, Lu Y, Yan Y. Effects of strain softening on the penetration resistance of offshore bucket foundation in nonhomogeneous clay. Ocean Eng
2019;193(2019). 106594: 1-16.
[45] Randolph MF, O’Neill, Stewart DP. Performance of suction anchors in fine-grained calcareous soils. In: Offshore technology conference. OTC 8831; 1998.
[46] Yao W, An H, Draper S, Cheng L, Harris JM. Experimental investigation of local scour around submerged piles in steady current. Coast Eng 2019;142:27–41. 15
pp.
[47] Yao W. An experimental study of scour around subsea structures. The University of Western Australia; 2019 [PhD thesis].
[48] Fu D, Zhang Y, Yan Y, Jostad HP. Effects of tension gap on the holding capacity of suction anchors. Mar Struct 2020;69(2020). 102679: 1-14.
[49] Fu D, Zhang Y, Aamodt KK, Yan Y. A multi-spring model for monopile analysis in soft clays. Mar Struct 2020;72(2020). 102768: 1-20.

22