Assessment on the potential use of vessel motion limit criteria for subsea cable installation

Abstract

Van Oord is active in the subsea cable installation industry and executes most of their cable lay projects with their cable laying vessel: The Nexus. In this thesis the focus lies on the normal lay phase of the cable installation, which comprises the phase when the vessel is pulling out the cable and putting it down on the seabed following the desired cable route. In order to ensure cable integrity during cable installation, a normal cable lay analysis is executed in the dynamic analysis software Orcaflex. The aim of the cable installation analysis is to define the installation limits in terms of the sea state. However, the vessel motions of the Nexus can be measured instantaneously and accurately on board of the vessel. Therefore, an assessment into the use of vessel motions for the expression of the handling limits during cable installation is performed. The assessment is executed for a normal lay configuration with an export cable and a 50 meter water depth. The focus lies on the maximum curvature and maximum tension response of the cable. The cable dynamics result from both the vessel motions and the direct cable loads. First, the effect of these phenomena is assessed independently with the use of Orcaflex simulations. The influence of the vessel motions on the cable dynamics is found to be low for short wave periods, as the vessel hardly reacts to these kinds of waves. As a result, vessel motion limit criteria are less suitable for expressing cable installation limits at less severe sea conditions. However, the handling limits of the cable are not likely to be exceeded during these kinds of sea states, therefore this does not directly prevent the use of vessel motion limit criteria. Next, the most suitable vessel motion, measured at the chute of the vessel, for application of vessel motion limit criteria is determined based on the time lagged cross correlations between the cable response and the vessel motions. By using this method, the vessel motion which is most linearly related to the cable response is selected. In the case study, the heave acceleration and axial acceleration of the vessel are identified as most suitable candidates for application of vessel motion limit criteria. Finally, the performance of the selected vessel motion as limit parameter is compared to the use of the wave elevation, equivalent to sea state limit criteria. The performance of both is assessed by a linear regression analysis of the peaks in the limit parameter time history and the associated peaks in the cable response. This analysis led to the conclusion that in the case study higher certainty can be given to vessel motion limit criteria compared to sea state limit criteria, which eventually can lead to an increase of the workability. Furthermore, a sensitivity analysis is performed to identify if the selected vessel motion for the application of vessel motion limit criteria is sensitive to changes in the normal lay configuration. The selected vessel motion and accompanied magnitude of the correlation were prone to changes in the normal lay configuration. Therefore, the applicability of vessel motion limit criteria for the base case in this thesis cannot straightforwardly be generalised for other normal lay configurations. Due to the nonlinear nature of the cable lay system, all cable installation analysis are executed using time domain simulations in Orcaflex, which are associated with large computational time. In light of reducing the computational time for normal lay analysis, the potential use of a transfer function for estimation of the maximum cable response is evaluated. The transfer function is set up based on the first order frequency response of the cable system to regular waves. Before application of the transfer function approach, the nonlinear behaviour of the system is studied on the basis of the spectral response of the cable towards regular wave simulations in Orcaflex. Especially the contribution of the higher order frequency components and the effect of the nonlinear drag term in the Morison equation are addressed. In order to check the performance of the cable response transfer function, the maximum cable response estimation of the transfer function for a three hour time duration is compared to the statistical three hour maximum resulting from non-linear Orcaflex simulations. The transfer function is found to underestimate both the curvature and the tension response of the cable, leading to the conclusion that this transfer function approach is not suitable for the prediction of the maximum cable response. The underestimation is caused by the high dynamic complexity of the normal lay system.