As turret moored units can weathervane against environmental parameters, heading angles play an important role for the design of FPSOs. The offshore structures need to withstand the most probable extreme responses of a target return period that the asset may encounter at an operational site. It is common that their structures are designed for mean heading angles obtained from heading analysis. However, it is a question whether the mean heading angles used for the long-term response calculations can represent the extreme event that the asset can encounter during the operation. In general, heading angle finds a new equilibrium angle due to weathervaning from environmental parameters. However, when environmental parameters have relatively large angles between them, equilibrium heading may not be stable enough but can show continuous slow yaw motions. Therefore, the variance of the heading angles was studied by time-domain dynamic analysis to be taken into account for long-term response calculations in this study. Due to the large amount of metocean data and required time for the dynamic analysis, ANN (Artificial Neural Network) regression technique was implemented to supplement the analysis. Several parameters were used to train the datasets and predict which sea state would lead to wider heading angles by ANN. Such results were compared to the long-term responses calculated with the mean heading angles in this study.
Floating offshore structures operating at fixed locations are designed to withstand site-specific loads. Classification societies have tailored their Rules and procedures for floating offshore units so that the site-specific environmental loads can be used for the structural design. Long-term responses of hull structures are derived statistically in the target return period, typically 100 year, to be used for either local scantling assessment or structural analysis. As turret-moored Floating Production, Storage, and Offloading (FPSO) units can weathervane against environmental parameters, a mean heading angle estimated for each sea state plays an important role in the determination of the design loads. The mean heading angles used for the short-term response calculations are normally determined by quasi-static equilibrium from various environmental parameters and this approach has been accepted by industry due to its fast calculation as a large number of sea states are to be processed for offshore projects. Milne et al. (2016) compared the quasi-static heading analysis method with full scale measurements of a FPSO in relatively benign sea conditions and showed good accuracy at stable time histories. This approach also provides good estimates of possible heading distributions of metocean data, however, it can neglect the events of lower probability by taking only a single mean heading for a sea state. Since, the design loads for local scantling, hull girder ultimate strength assessment and strength analysis should be based on the most probable extreme response, all the probable events need to be taken into account if they are still likely to occur during the operation.
