2nd Order Hydrodynamic Effects on Resonant Heave, Pitch and Roll Motions of a Large-Volume Semi-Submersible Platform

Abstract

During the last decades, as oil production offshore Brazil moved to deeper waters, technical and economical constraints led to a new generation of floating platforms. Nowadays, in the Brazilian offshore scenario, design trends concerning hull form, size and mooring configurations bring novel characteristics of wave-induced dynamics, including non-linear resonant effects. As part of an extensive study on new semi-submersible configurations for Campos basin, recent model tests have shown that their hulls may be subjected to second-order slow motions in heave, pitch and roll. These resonant motions are directly related to the large dimensions and relatively low natural frequencies of the floating systems. The unexpected effects caused great concern, since, in some cases, the low-frequency motions presented amplitudes comparable to those of the first-order response. This paper discusses the evaluation of the 2nd order wave-induced motions of a large-volume semi-submersible platform using WAMIT® second-order module. It is shown that the hydrodynamic forces induced by the 2nd-order potential represent the prevailing effect in the resonant response. Important aspects concerning the numerical model are addressed, such as the parameters involved in the hull and free-surface panelization. Numerical predictions are directly compared with experimental results obtained with a 1:40 model of the platform. A very good agreement is obtained both for heave and angular (pitch or roll) motions, attesting that the numerical code is able to predict the 2nd order forces accurately. Finally, a simplified procedure for dealing with the slow vertical motions is evaluated, aiming to reduce the substantial computational effort required by the 2nd order calculations. Such procedure takes advantage from the fact that the resonant response spectra of the vertical motions are usually narrow-banded (due to the low damping levels) to propose a “white-noise” approach. According to this approach, 2nd order forces need to be calculated only for one frequency difference, corresponding to the natural frequency of the particular motion. Computational time is, therefore, greatly reduced. It is shown that resonant motions calculated through the simplified approach match those predicted through the “full” analysis perfectly, making it an interesting choice for the evaluation of 2nd order effects, especially in the early stages of the design.