*1.1. Background*

Jack-up units are self-elevating mobile platforms which are used extensively in the offshore oil and gas industry. A typical jack-up consists of a floatable hull and three independent retractable legs. The legs rest on spudcan footings that are usually circular or polygonal in plan and with an inverse cone underneath. Once a jack-up unit is towed to site, its installation begins by lowering the legs to the seabed and pushing the spudcans into the soil and then rising the hull over the water. Then pre-loading can be achieved by pumping water into the hull. The pre-loading makes the spudcan penetrate deeper to provide more resistance. After pre-loading, the water is pumped out and the spudcan's bearing capacity has some reservation. After all the work of the jack-up has finished, it is removed from the site by retracting the legs from the seabed. The processes of installation and extraction of the spudcan leave a permanent seabed depression at each footing site, which is referred to as a "footprint".

The footprint changes the seabed in two ways, as shown in Figure 1: An inclined seabed surface and a varying soil strength profile within the footprint (normally decreasing soil strength due to remolding). Both of them result in additional horizontal forces and bending moments compared with the initial installation. The spudcan–footprint interaction problem is significant as it can lead to significant time loss, cost implications, risks to adjacent structures and potential injury to personnel. Dier et al.

concluded from industry practice data that incidents caused by uneven seabed/scour/footprint are at a rate of 15% of the total [1]. This rate has increased obviously due to increasing demands of jack-up operation close to previous sites in recent years [2]. significant time loss, cost implications, risks to adjacent structures and potential injury to personnel. Dier et al. concluded from industry practice data that incidents caused by uneven seabed/scour/footprint are at a rate of 15% of the total [1]. This rate has increased obviously due to increasing demands of jack-up operation close to previous sites in recent years [2]. seabed/scour/footprint are at a rate of 15% of the total [1]. This rate has increased obviously due to increasing demands of jack-up operation close to previous sites in recent years [2].

*J. Mar. Sci. Eng.* **2019**, *7*, 175 2 of 19

the initial installation. The spudcan–footprint interaction problem is significant as it can lead to

*J. Mar. Sci. Eng.* **2019**, *7*, 175 2 of 19

the initial installation. The spudcan–footprint interaction problem is significant as it can lead to

Dier et al. concluded from industry practice data that incidents caused by uneven

**Figure 1.** The failure mechanism of spudcan reinstallation near a footprint. **Figure 1.** The failure mechanism of spudcan reinstallation near a footprint. *1.2. Previous Work* 

#### *1.2. Previous Work 1.2. Previous Work* The problem of jack-up reinstallation near the existing footprints attracted more attention in the

The problem of jack-up reinstallation near the existing footprints attracted more attention in the recent 10–15 years. In some studies [3,4], footprints resemble an inverted conical shape cavity. The installation, operation, and removal of the spudcan can also remold the surrounding soil, resulting in highly variable shear strength profiles in the vicinity of the footprints [4–8]. The problem of jack-up reinstallation near the existing footprints attracted more attention in the recent 10–15 years. In some studies [3,4], footprints resemble an inverted conical shape cavity. The installation, operation, and removal of the spudcan can also remold the surrounding soil, resulting in highly variable shear strength profiles in the vicinity of the footprints [4–8]. recent 10–15 years. In some studies [3,4], footprints resemble an inverted conical shape cavity. The installation, operation, and removal of the spudcan can also remold the surrounding soil, resulting in highly variable shear strength profiles in the vicinity of the footprints [4–8]. Hartono et al. [9] used an experimental method (centrifuge tests) and numerical analysis (simulated with ABAQUS/CEL) respectively to investigate the efficacy of reaming technique in

Hartono et al. [9] used an experimental method (centrifuge tests) and numerical analysis (simulated with ABAQUS/CEL) respectively to investigate the efficacy of reaming technique in mitigating the footprint hazards. He found that the numerical results demonstrate good agreement with experiment results and reaming can be a viable option to mitigate spudcan–footprint interaction. He strongly suggested making numerical modeling as a viable tool for site-specific assessment of spudcan–footprint interaction problem. Like Hartono, the CEL large deformation method is adopted in this study to investigate the reinstallation behaviors of flat base footing, fusiform spudcan footing, Hartono et al. [9] used an experimental method (centrifuge tests) and numerical analysis (simulated with ABAQUS/CEL) respectively to investigate the efficacy of reaming technique in mitigating the footprint hazards. He found that the numerical results demonstrate good agreement with experiment results and reaming can be a viable option to mitigate spudcan–footprint interaction. He strongly suggested making numerical modeling as a viable tool for site-specific assessment of spudcan–footprint interaction problem. Like Hartono, the CEL large deformation method is adopted in this study to investigate the reinstallation behaviors of flat base footing, fusiform spudcan footing, and skirted footing. mitigating the footprint hazards. He found that the numerical results demonstrate good agreement with experiment results and reaming can be a viable option to mitigate spudcan–footprint interaction. He strongly suggested making numerical modeling as a viable tool for site-specific assessment of spudcan–footprint interaction problem. Like Hartono, the CEL large deformation method is adopted in this study to investigate the reinstallation behaviors of flat base footing, fusiform spudcan footing, and skirted footing. Spudcans are the most common footings used for jack-up units. Along with the improvement of

and skirted footing. Spudcans are the most common footings used for jack-up units. Along with the improvement of technology and the increasing demands of operating on the very soft soils, the footings become larger in diameter and flatter at the base. The geometries of typical fusiform spudcan footings are shown in Figure 2. The investigations from some research shows that, by comparing with fusiform spudcan footing, skirted footing may have a higher bearing capacity [7] and have some potential in mitigating Spudcans are the most common footings used for jack-up units. Along with the improvement of technology and the increasing demands of operating on the very soft soils, the footings become larger in diameter and flatter at the base. The geometries of typical fusiform spudcan footings are shown in Figure 2. The investigations from some research shows that, by comparing with fusiform spudcan footing, skirted footing may have a higher bearing capacity [7] and have some potential in mitigating punch-through failure [10,11]. technology and the increasing demands of operating on the very soft soils, the footings become larger in diameter and flatter at the base. The geometries of typical fusiform spudcan footings are shown in Figure 2. The investigations from some research shows that, by comparing with fusiform spudcan footing, skirted footing may have a higher bearing capacity [7] and have some potential in mitigating punch-through failure [10,11].

Mod V "A"(Cassidy et, al 2009) 116C jack-ups (Cassidy et, al 2009) HYSY 944 (Hai-bo, Y. 2016) **Figure 2.** Three typical spudcans[3,12]. **Figure 2.** Three typical spudcans [3,12].

**Figure 2.** Three typical spudcans[3,12]. Cassidy et al. [3] used a 1:250 scale model of current Mod 'V 'jack-up in their centrifuge tests to simulate the interaction between real spudcan and soil. Kong et al. [13,14] replaced real spudcan with flat base footing in their centrifuge tests to eliminate the variables related to spudcan geometry. Cassidy et al. [3] used a 1:250 scale model of current Mod 'V 'jack-up in their centrifuge tests to simulate the interaction between real spudcan and soil. Kong et al. [13,14] replaced real spudcan with flat base footing in their centrifuge tests to eliminate the variables related to spudcan geometry. Cassidy et al. [3] used a 1:250 scale model of current Mod 'V 'jack-up in their centrifuge tests to simulate the interaction between real spudcan and soil. Kong et al. [13,14] replaced real spudcan with flat base footing in their centrifuge tests to eliminate the variables related to spudcan geometry. Zhang et al. [15] simulated spudcan with flat base footing in his numerical study to make sure that thetouchdown level of footing could be identified clearly.

Gan et al. [4,6,8] studied the spudcan–footprint interaction considering the 'real' initial penetration. Their research showed that the soil is obviously disturbed during the initial penetration and will recover with time. To simplify the problem, many of the following studies assumed an artificial footprint, such as Kong [13], Zhang et al. [15], Jun et al. [16–19]. The assumption of an artificial reverse cone footprint may respond to a fully recovered 'real' footprint after a long period from the initial penetration. Therefore, the three idealized footprints TA, TB, and TC following Kong [13] are adopted in this study to simplify the numerical model.

The distance from the central line of the footprint to the reinstalled spudcan center was termed as offset distance or reinstallation eccentricity (β), which was proved to be a key issue to the profiles of bending moment (*M*) and horizontal force (*H*). Stewart [20] carried out centrifuge model tests and the results showed that both *M* and *H* increased to an obvious value when β/*D* = 0.5 to 1.0, where *D* is the diameter of the reinstalled spudcan, and *H* reached to the maximum value when β/*D* = 0.75. Cassidy et al. [3] founded that *M* and *H* were most obvious when β/*D* = 0.5 and became very small when β/*D* > 1.5. Carrington [21] carried out large deformation numerical analyses to simulate the reinstallation processes with β/*D* = 0.167 to 0.407, and obtained a most critical case at β/*D* = 0.29. Kong et al. [13,14]. Investigated the effect of footprints with various size and slope angles. In their study the critical case was β/*D* = 1.0.

Some research showed that the fixity condition at the leg–hull connecting point has a significant effect on the reinstallation behavior near a footprint [3,5,22–25]. It can be concluded that harder fixity tends to increase the maximum value of *M* and *H* but reducing the lateral movement of the spudcan during reinstallation.
