Experimental modeling of pile-leg interaction in jacket type offshore platforms cyclic inelastic behavior

American Journal of Applied Sciences, Nov, 2008 by M.R. Honarvar, M.R. Bahaari, B. Asgarian

INTRODUCTION

General: Most of the jackets have foundation piles through each main leg which are welded to the structure at deck level. Modeling the cyclic inelastic behavior of legs and braces is of utmost importance in any nonlinear dynamic analysis of offshore jackets. In the past years, considerable investigative efforts have focused on understanding the inelastic behavior of offshore structures subjected to severe seismic loadings and numerical methods for predicting this behavior have been developed (15). Numerical efforts on frame behavior used to be evaluated by experimental tests which were all conducted in a simplified condition without modeling the pile above seabed and pile-leg structural interaction.

In this study, two scaled 2Dimenional models of a platform are fabricated and tested under cyclic deck displacement to show the effect of pile-leg interaction in inelastic range of deformation. Furthermore the different behavioral aspects of grouted and un-grouted pile-leg interaction are investigated.

Where a structure's piles pass through the legs, the primary pile/structure connection is typically made by welding the pile to the structure and/or by grouting the leg/pile annulus. In this case, the pile leg acts as a composite member. In case of un-grouted leg-piles, the effect of shim plates at elevation of horizontal braces have to be considered (10).

Background and scope: Zayas et al. (14) reported the results of a series of axial cyclic tests on isolated struts representing typical tubular braces in offshore platforms. Zayas et al. (15), reported the experiments on two one-sixth scale models of an X-braced tubular steel offshore platform subjected to cyclic displacements simulating the effects of severe earthquake ground motions. These efforts were numerically studied by Keyvani (11) and Asgarian (4). Ray Clough and Yousof Ghanaat (8) have studied the dynamic elastic and inelastic behavior of one 5/48 scaled model of an X-braced offshore platform made of tubular members. In this platform, a 5/8 scaled model of the jacket tested by Zayas et al. was represented. Earthquake motions used for jacket shaking were modified records using El Centro and Taft recording earthquakes. Linear and nonlinear analyses of the structure on the shaking table were performed using fiber Beam-Column post-buckling element by Asgarian (3).

In this study, a complementary test is conducted in line with Zayas et al. efforts (15) to model the pile-leg real interaction through the connection joint and shim-plates and show the effect of grouting in lateral cyclic behavior of offshore jacket type structures.

The experimental program and the frame models employed are described in section 2 of this paper. Two types of frames, representing the current practices are investigated. In one model, the gap between pile and leg is filled with grout, which is called Grouted frame and the other frame is called Un-grouted. The overall behavior of the experimental model is discussed in section 3 of this paper. Particular attention is paid to the effect of grouting on the deterioration of strength and energy dissipation capacity.

Jacket members behavior in nonlinear range of deformation

Portals: Portals are members with essentially constant axial force and variable lateral displacement (11). The portal behavior of a jacket leg under lateral forces results in a relatively small bending moment at the mid-height of the leg segment between any two adjacent horizontal bracing levels. Grouted and Un-grouted portals behave differently and have different methods of modeling (10).

Struts: Struts are essentially bracing members with constant lateral loads (usually zero) but variable axial displacement (11). Since jacket braces are connected to legs, which normally have much larger flexural stiffness, their axial response is their most important behavioral characteristic.

Quasi-static loading approach: A good approach to verify specific aspects of the numerical model, as was utilized here, is to impose quasi-static load or deformation histories on the specimen. This type of study provides valuable information on force-deflection relationship of the frame as a whole and also on individual members. Realistic numerical modeling of these loops is essential for accurate computer simulation of the inelastic dynamic response of a structure. Past research on braced frames has been also mostly performed in a quasi-static manner (15). If it is desired to obtain information to develop and verify numerical models which can subsequently be used to predict dynamic behavior, it is not necessary to perform tests dynamically, because:

* Once a numerical model is verified on the basis of experimental data under cyclic loading, time history dynamic analysis can be used with confidence to check ductility demand, the structural displacements and the energy dissipation of the structure (15)

* Earthquake excitations are non-deterministic and several records must be applied to assess the nonlinear structural response properly (15)