Frequency and Time Domain Analyses of Vessel Motions During Tropical Cyclones

A permanently turret moored floating facility located off the North West coast of Australia is likely to be exposed to tropical cyclones, which exhibit extreme wave height, wind and surface currents. Furthermore the cyclonic eye and fringes are characterised by rapidly varying metocean conditions, particularly wind speed and direction. It is necessary to understand the weather vaning and motion responses of the vessel during these transient conditions for successful facility development. This paper presents weather vaning and motion analyses of a Floating Storage and Offloading (FSO) vessel during the passage of tropical cyclones. A synthetic tropical data base developed by Woodside Energy Limited was used to generate the cyclonic conditions. The data base contains storms with return periods ranging from 10 to 100,000 years. Time histories of wave spectrum, wind and current at half-hourly intervals for several cyclones were examined. The weather vaning and motions responses of the FSO were computed using two methods: (1) a frequency domain quasi-stationary approach, and (2) a time domain approach. In the frequency domain analyses the metocean conditions were treated as stationary for each half-hour interval and the mean vessel heading and most probable amplitudes of motions were calculated. In the time domain analyses, the time histories of metocean conditions were closely matched to the synthesised cyclonic conditions, and time histories of vessel heading and motion responses were generated. Multiple realisations were simulated for each cyclone to assess variability of results associated with wave train random seed. A key finding of the study is that the worst roll response tends to occur after the eye of the cyclone has passed, at which time the wind and waves were highly non-collinear. At this time the vessel weather vanes so as to experience waves which are substantially beam-on. When the cyclonic eye is directly over the site, there is a reduction in wind speed and as a consequence the vessel weather vanes into the sea. This causes a corresponding reduction in the roll response. Results from both analysis methods have been compared and good agreement is observed. The applicability and limitations of the two analyses methods are also considered.