Mobile jack-up drilling rigs often need to return to a site where a previous installation has left footprints in the seabed. Reinstallation near these depressions is a problematic operation because the jack-up's circular spudcan footings become subjected to eccentric and (or) inclined loading conditions. This can lead to structural failures within the jack-up legs and (or) excessive leg tilt and hull displacement. This paper reports a comprehensive set of geotechnical centrifuge experiments that investigated the effect of footprint geometry on the reinstallation response. Artificial conical shaped footprints were manually cut in the centrifuge sample, ensuring consistent shapes and minimizing any variation of undrained shear strength due to the process of initially installing and retrieving a spudcan. The effect of footprint geometry was thereby isolated. The vertical, horizontal, and moment loads induced on a model footing when penetrated at varying offsets are presented and these provide evidence on the effect of different footprint depths and angles on installation. The footprint geometry governed the horizontal force and moment observed during reinstallation between the level of the touchdown and the footprint toe. Further experimentation has shown that an equivalent skirted footing induced significantly higher horizontal forces (although it can be assumed to be significantly stiffer).

The reinstallation of jack-up units near a pre-existing footprint is one of thechallenges currently faced by the jack-up industry. Footprints have an unevensoil surface and heterogeneous strength characteristics. As a consequence, thespudcan re-installation is resisted by an eccentric and/or inclined reactionfrom the soil. This results in spudcan and leg tilting that is in turn resistedby the development of a bending moment in the leg, which may compromise thestructural integrity of the jack-up unit. The paper presents an overview of the research performed at UWA on footprintspudcan interactions, highlighting (i) the relative contributions of thefootprint geometry and soil heterogeneity to the development of bending momentsin the spudcan leg and (ii) the necessity of modelling the spudcan and legfixities correctly to assess the forces generated in the spudcan and leg duringreinstallation accurately. Introduction Jack-up units are commonly used for oil or gas exploitation in water depths upto 120 m. They consist of a buoyant triangular hull connected to threeindependent truss-work legs with a conical shape foundation (known as aspudcan) at the base. During installation, the legs are lowered into the seabedindependently, usually one after another. The loading process includes stagesof preloading where additional load (in comparison to the in-service load) isapplied on the spudcan. Once the drilling or work over is complete, the jack-upunit is removed, leaving footprints on the seabed, which may be up to 10 m deepand 30 m wide in soft clay. The reinstallation of jack-up units nearby pre-existing footprints is one ofthe challenges currently faced by the offshore oil and gas industry (Figure 2). During the installation process, vertical load is applied directly though thecentre of the spudcan. In the case of installation near a pre-existingfootprint, where the soil surface is uneven and may exhibit heterogeneousstrength, an eccentric and/or inclined reaction from the soil may be applied tothe spudcan, which will tend to cause tilting of the spudcan and lateral motionof the leg. These motions are resisted by the development of a bending momentin the leg and at the hull-leg connection, potentially threatening thestructural integrity of the jack-up leg. Although this problem has been clearly identified, there are still noguidelines to assist operators in a safe reinstallation, aside from therecommendation to monitor leg loads via the rack phase difference duringinstallation and to create an even soil surface over footprints. There isconsequently a need to understand the parameters governing thespudcan/footprint interaction and to be able to predict forces generated in thespudcan leg during reinstallation. Reinstallation scenario and problem relevance Due to their mobility, jack-up units may be moved relatively easily from onelocation to another before coming back to an existing location to drilladditional wells or to enhance the production of existing wells. During theextraction process, each leg leaves a depression on the seabed, known as afootprint, of highly disturbed strength, and with geometry varying with thesize and shape of the penetrating spudcan and the characteristics of the soil. Figure 1a presents the result of a seafloor survey undertaken after jack-upremoval that highlights the presence of footprints. A statistical studyundertaken by Berg (2004) indicated that there were approximately 1200footprints within the Shell EP Europe at the time of the survey.

No abstract

This paper presents a pioneering real-time hybrid testing method for geotechnical centrifuges. The method was used to investigate the behaviour of a jack-up leg reinstalled near an existing footprint, a problem that is highly nonlinear, stress dependent and involves complex soil-structure interactions. By physically modelling only the footprintleg interaction and numerically modelling the rest of the jack-up structure, the method enables a realistic account of all the parameters involved in the interaction, including the footprint geometry, soil heterogeneity and structural properties of the jack-up unit. The paper also details the three-degree-of-freedom actuators developed to model the interaction, which features controlled loads and motions along the vertical, horizontal and moment directions, and the real-time algorithm that bridges the physical and numerical models. The algorithm allows the stiffness on each axis to be varied. Testing performed at 1g and in the centrifuge, modelling jack-ups with different stiffness on the horizontal and moment axes, is presented to validate the apparatus and methodology. The hybrid apparatus and real-time testing method were found to produce much more realistic boundary constraints than previous fixedsystem apparatuses, and this allowed the test results to be considerably more informative, accurate and useful.

No abstract