Dynamic behaviour of Kvitebjorn jacket structure - Numerical predictions versus full-scale measurements

The dynamic response of the Kvitebjorn jacket is considered. Full scale measurements are compared with time domain simulations utilizing the design model. Topside weight and soil-structure stiffness, however, have been modified in order to match observed natural periods. The platform is a slender steel structure installed in about 190m water depth in the Northern North Sea in the summer 2003. The wave induced loads on this type of platform are for storm sea states governed by the non-linear drag term of the Morison equation. Due to this type of non-linearity, and the fact one has to integrate the drag load density to the exact surface in order to obtain the global drag loading correctly, the structure will experience excitation forces both at the wave frequency, omega, and at multiples of this frequency, 2 omega,3 omega,.... In design it was expected that the 3 omega term could result in a considerable dynamic amplification in connection with the largest waves since it would cause forcing at a period close to the largest natural period which was estimated to be 5s under design assumptions. Platform measurements have shown that the largest natural period is significantly lower. Around 4s is estimated during the first winter. Reasons for this will be discussed. The consequences of the lower period on predicted response will be indicated, but the major emphasis of this paper is to compare time domain simulations with observed response when the measured wave process is taken as input to the simulations. An important observation is that the 3 omega load term is less important than feared in the design process even after adjusting platform properties to those matching observed natural periods.