Jackup Rig Foundation Modeling II

ABSTRACT As part of an on-going research effort, drum centrifuge tests of three-leg jack-up rigs on sand and clay seabeds have been used to improve an elastic-plastic work-hardening model for jack-up foundations previously proposed [6]1. This paper constitutes an interim research progress report. The spud can secant fixity for sand was found to vary from 0.6 to 0.9 (on a scale of 0.0 to 1.0 proposed herein) and is consistent with field measurements [3]. The inferred secant elastic rotational modull for loose and dense sands differ by a factor of 1.3 to 1.7 and both are significantly greater than those determined in single-leg tests [6]. The secant fixity and modull for clay during 2-way cyclic loading degrade to asymptotic values within the first 10 to 20 loading cycles. The relative magnitudes of modulus degradation are consistent with previous findings [4], which were based on field measurements. However, the effects of cyclic loading on tangent fixity and modulus are inconclusive. A simple hyperbolic model with asymptotic behavior based on theories [1, 2] only slightly underpredicts the interpreted moment-rotation response of spud cans in clay. INTRODUCTION The response of an independent leg jack-up to environmental loading is highly sensitive to foundation behavior, especially the rotational "fixity" of the spud can. Two key response variables, deck-leg moment and natural period of the structure, are directly affected by the foundation fixity condition. To investigate this problem experimentally, a geotechnical drum centrifuge, which provides properly scaled model tests of three-leg jack-up rigs subjected to lateral loading, was used. The results reported herein are intended to extend previous work on this topic [5, 6] and to provide an interim progress report on an ongoing experimental research program. In this study, the results of extensive drum centrifuge tests on three-leg jack-up rig models supported on uniform sand (loose and dense) and clay seabeds are examined. In particular, the sub-yield behavior of the spud can rotational response is extracted, which may serve as a rough guideline in the selection of "elastic" input parameters to the elastic-plastic work-hardening foundation model previously proposed [6]. Note that the "elastic" properties represent idealized behavior and are not truly elastic. The elastic rotational response due to horizontal loading is interpreted in terms of spud can fixity, which is the ratio of the actual moment at the mudline to the theoretical moment of a spud can that is fully fixed against rotation. The rotational stiffness (and hence "elastic" modulus) can then be determined since it is a unique function of fixity, The use of spud can fixity to estimate the elastic rotational modulus circumvents the uncertainties associated with the use of very small spud can displacements (fractions of a millimeter) inferred from transducer measurements, which was a drawback with single- Ieg tests [5, 6, 7]. The inferred modull for loose and dense sands are compared with the empirical relationships derived from single-leg test results [6, 7].