Coupled Analysis of a Moored Sevan Hull by the use of OrcaFlex

Abstract

Offshore activity related to the search of energy in terms of hydrocarbons is taking place at increasing water depths. In deeper water, the effects on the floater motion from the mooring and riser systems increase significantly. Viscous damping, inertial mass, current loading and restoring effects from riser and mooring systems need to be adequately modeled in order to predict the system motion of response. Accurate prediction of the system response is vital in the context of design verification of the complete system. The recommended method to analyze the dynamics of a moored structure, especially in deep waters, is by a coupled analysis approach. This is recognized by the dynamics of the floater, mooring lines and risers being solved simultaneously. This thesis comprises a theoretical part, where key subjects related to the prediction of moored structure motion are addressed. In addition, coupled time domain analyses of a moored Sevan FSU (Floating Storage Unit) were performed. The analyses were initiated by the establishment of a finite element model of the Sevan hull. This was used to conduct a diffraction analysis on FSU. The results obtained from the diffraction analysis were imported to OrcaFlex, where the mooring systems were also modeled. OrcaFlex is a fully 3D non-linear time domain finite element program, which supports Newman?s approximation and a method based on Aranha?s approximation in order to predict the low frequency motion of moored structures. Model tests have been previously performed on the moored FSU, and results from these tests were used as the basis for comparison to the OrcaFlex results. In addition, time series of the wave elevation and wind velocities recorded during the FSU model tests were acquired through Sevan and Marintek, enabling these environmental conditions to be integrated in the OrcaFlex analyses. Hence time series results from the OrcaFlex analyses were directly comparable with the FSU model test time series.The drawback of this approach ultimately proved to be the CPU-cost and the duration of the numerical simulation of relatively short model test time series (scaled to real time). The results obtained from the OrcaFlex analyses show acceptable compliance to the model test results, but the amplitude of the low frequency motion seems to be slightly under predicted. In order to predict outcomes under specific scenarios e.g. extreme events of the system response, a longer real time simulation is required.