Experimental behavior of piggy-backed anchors in cohesionless soils

As the offshore wind platform ventured into deep water, improving anchor efficiency has become the main focus of foundation design. Plate anchors are attractive options due to factors such as high anchor efficiency and low cost. Appropriately attaching two plate anchors improves the total capacity of the piggy-backed anchors configuration, and generally, this capacity is greater than twice the capacity of an individual anchor. This paper presents the centrifuge drag embedment tests of an individual drag embedment anchor (DEA) and the piggy-backed DEAs in cohesionless soil. The piggy-backed configuration with the spacing between the two anchors at three fluke lengths and an attachment point at the back of the fluke can improve the anchor total capacity. The main reason is that a deeper penetration of the back anchor and the steeper pitch of the front anchor facilitate both anchors to dive deeper.


Introduction
Plate anchors are widely employed in offshore foundations due to the high anchor efficiency, which is defined based on the holding capacity in relation to the foundation's self-weight.Among these plate anchors, the drag embedment anchor (DEA) stands out given that it is characterized by a bearing plate (referred to as a fluke) and a series of plates or lines (referred to as a shank), as depicted in figure 1.The installation of a DEA involves initially positioning the anchor at an appropriate fluke-shank angle based on the in-situ soil type.Subsequently, the anchor rope or chain is pulled to drive it downward into the seabed.The setting of the fluke-shank angle typically ranges around 30° in sands and approximately 50° in soft clays.Despite the prevalent use of DEAs in sands and stiff clays, these anchors consistently encounter shallow embedment, which is often less than 1-2 fluke lengths (Aubeny, 2018).
The drag embedment anchor (DEA) performance has been explored through the integration of a piggy-backed anchors concept, which involves the attachment of two DEAs in series (see figure 2).This configuration utilizes existing installation equipment and lightweight anchors, thus offering a promising avenue for enhancing anchor embedment in sand.Notably, 1g drag embedment model tests in sand have demonstrated that a piggy-backed arrangement, which is characterized by an appropriate attachment point and spacing between the two anchors exceeding 2 fluke lengths, facilitates anchor penetration effectively [1].Consequently, the collective capacity of piggy-backed anchors is substantially higher, at least twofold, compared to that of an individual anchor.Conversely, linking the rear anchor to the palm and shank of the front anchor induces roll instability.Field tests conducted by Taylor [2] using two STATO or NAVMOOR anchors at Port Hueneme, featuring well-grained fine sand, affirmed that a judiciously configured piggy-backed arrangement facilitates anchor embedment and may achieve equivalent or superior anchor capacity compared to utilizing two lightweight anchors.This accumulating 1337 (2024) 012048 IOP Publishing doi:10.1088/1755-1315/1337/1/012048 2 body of evidence indicates that piggy-backed anchors possess the potential to enhance embedment, particularly in sands exhibiting relatively high resistance.
The primary objective of this paper is to provide valuable insights into the behavior of piggy-backed drag embedment anchors (piggy-backed DEAs) in cohesionless soils.The methodology involves conducting centrifuge tests with both an individual DEA and piggy-backed DEAs.Through a comprehensive analysis of trajectories, capacities, and pitch, we aim to elucidate the superior capacities attained by piggy-backed anchors.

Experimental Setup 2.1. Test bed
The soil utilized in this study consisted of Fujian standard sand, characterized by an average particle size (d50) of 0.17 mm.Notably, the ratio of model anchor width to average particle size (df / d50) exceeded 40, rendering the impact of soil particle size negligible as per the findings of Craig [3].The soil container, with internal dimensions measuring 1.2 m in length, 0.95 m in width, and 1.0 m in depth, was prepared for experimentation.To facilitate proper drainage, a 0.05-m-thick initial layer of coarse sand was placed at the bottom of the container, overlaid with a nonwoven geotextile.Subsequently, dry sand was meticulously poured using a spot pouring hopper, achieving a thickness of 0.5 m and a relative density of 60%.The soil container was hermetically sealed within a vacuum box for one week, during which air-free water was introduced through the base of the model box to ensure complete saturation of the sand.The key soil properties are presented in table 1.

Magnetometer
The magnetometer device employed in this study is the Micro Sensor 1.8 TM , a product from the Patriot series manufactured by Polhemus.This device comprises an electronic system unit, a magnetic source, and a microsensor, as shown in figure 4. The electronic system unit oversees the magnetic source, directing the excitation of the magnetic field while simultaneously processing the signal data from the IOP Publishing doi:10.1088/1755-1315/1337/1/0120484 sensors, as illustrated in figure 5.The magnetic source itself is a three-axis orthogonal coil enclosed within a 5.84-cm² cube box, which generates an electromagnetic field.The microsensor, which possess a diameter of 1.8 mm and is also configured as an orthogonal three-axis coil, is responsible for generating receiving signals.Notably, it can be easily affixed to the surface of the offshore anchor model without introducing any impact on the test results.

Centrifuge test
The ZJU-400 centrifuge, illustrated in figure 6, located at Zhejiang University, boasts an effective arm radius of 4.5 m, a maximum centripetal acceleration of 150 g, and a formidable capacity reaching 400 g•ton.The swinging basket affiliated with this centrifuge possesses dimensions of 1.5 ×1.2 ×1.5 m.

Figure 6. ZJU-400 centrifuge
Figure 7 illustrates the setup of the DEA installation in the swing platform shown in figure 6.A variable frequency motor is employed to pull the anchor through a pulley at a constant loading rate of 1 mm/s.The initial buried depth of the anchor model is set at 0 Lf with a drag distance of approximately 5 Lf.The initial loading angle of the anchor line at the mudline is fixed at 2.5° and is incrementally adjusted until it reaches 18.5° as the anchor gradually approaches the loading pulley.To capture the

Individual anchor
In figure 8, a noteworthy alignment is observed between the predicted trajectory, pitch, and tension, and the corresponding experimental findings.This alignment holds true across varying loading angles at the mudline during installation.The ultimate embedment depth of the DEA tip is consistently less than 1 Lf, a result that is in harmony with multiple model tests [6] (see also Luger and Harkes 2014).Upon reaching the ultimate embedment depth, as the drag distance and loading angle at the mudline increase, the fluke undergoes an upward reorientation.Consequently, the drag anchor initiates an upward extraction, leading to a subsequent decrease in bearing capacity.

Conclusion
Accurate measurement of the drag trajectory during installation, which encompasses drag distance and embedment depth, is imperative for assessing the load capacity of drag anchor installations.study employs a six-degree-of-freedom magnetic positioning system to conduct centrifugal model tests on drag anchor installations in sand.Both the individual anchor and the piggy-backed anchors with configuration of attachment point at the back of fluke and spacing between the two DEAs is 3 Lf.The piggy-backed anchors can yield a capacity that is twice the capacity of an individual anchor.The deeper penetration of the back anchor and the steeper pitch of the front anchor facilitate both anchors to dive deeper.

Figure 1 .
Figure 1.Installation process of the drag embedment anchor.

2. 2 Figure 3 .
. Anchor The model anchor was derived from a modified version of the simplified DEA with a fluke-shank angle set at 35°, as originally designed by Delmar Systems, Inc. Constructed from stainless steel, the model anchor was scaled at 1:30 in comparison to the prototype.The submerged weight (Wa') of the model anchor amounted to 212 g, with a volume (Va) of 30,690 mm³ .Key dimensions of the anchor included fluke length (Lf) of 91 mm, width (df) of 134 mm, and thickness (tf) of 3 mm.It featured a fluke area (Af) of 6,960 mm² , and a vertically projected shank area (As) of 328 mm² .In adherence to typical commercial anchor specifications, the length to fluke thickness ratio generally falls within the range 5-30; the present model aligns with this standard at 29.The fluke width to fluke length ratio typically varies from 1 to 2, and the model anchor conforms to a ratio of 1.5.The anchor line, which is fabricated by braiding highstrength PE (ultra-high molecular weight polyethylene fiber, UHMWPE) fishing line, possessed a diameter of 2 mm.Distinct anchor line types can be characterized by varying values of the parameter   / 2 , with 1300 representing typical wire line anchor systems and 150 indicative of a typical chain system [4].In the present test, the selected   / 2 is 1740.The piggy-backed anchors employed in this study involve attaching two DEAs with spacing at 3 Lf and the attachment point is at the back of the fluke.Views of model anchor: (a) side; (b) front; (c) bottom; and (d) top.

Figure 5 .
Figure 5. Operating principle of the magnetometer system.

Figure 7 .
anchor line, a load cell with a 1-kN capacity is vertically positioned between the two pulleys of the pulley frame.Kinematic data in the six degrees of freedom of the anchor (x, y, z, ψ, θ, and φ) are recorded using the magnetometer system.A Plexiglas bracket secures the magnetic source to the center of the model box, ensuring a distance of over 300 mm from the aluminum model box wall.This setup guarantees measurement accuracy for the magnetometer (Lai, 2022b).Centrifuge tests setup: (a) sketch for test setup; and (b) picture for measuring system.
Figure 9(a) indicates that the capacity of the piggy-backed DEAs may be enhanced by attaching the back anchor to the back of fluke of the front anchor.Figure 9(b) and figure 9(c) presents the corresponding trajectory and pitch, respectively.Figure 9(b) shows the attachment point at the back of fluke facilitates the back anchor dive deeper, while the penetration depth of the front anchor is generally the same an individual anchor.In general, a deeper embedment depth results in a higher capacity.This is consistent with the greater capacity for attachment point at the back of fluke observed in figure 9(a).

Figure 9 .
Figure 9(c) shows that the front anchor possesses a steeper pitch angle while the back anchor has a flatter angle.A typical anchor behavior is that a deeper penetration potentially occurs upon a steeper pitch angle.Moreover, the capacity is expected to increase further with additional drag.This is reasonable given that the steeper pitch angle of the front anchor produces it to dive deeper, and the front anchor drives the back anchor (already deeply embedded) penetrates a greater depth.Therefore, this reflects the piggy-backed DEAs with the configuration of the attachment point at the back of fluke can yield a greater capacity.Test results for piggy-backed anchors: (a) trajectory; (b) pitch; and (c) tension.