LAUSR

Abstract As one of the key met-ocean environmental loads affecting underwater structural design, current is closely related to the failure process of underwater structures such as moorings and risers. In marine structural analysis, a current model depends mainly on its speed profile corresponding to the design criterion. The conventional model of the current speed extreme (CSE) profile ignores the interlayer correlation and renders the current profile model conservative. This study is focused on the analysis of a current profile model aimed at the strength and dynamic design of underwater structures with special consideration of the interlayer correlation based on the long-term prototype measured data in the “Liuhua” oil field in the South China Sea. For the current model related to strength design, an empirical orthogonal function (EOF) decomposition is first utilized for dimensional reduction. Then, the multi-dimensional extreme-value problem is converted into an inverse reliability problem. The inverse second-order reliability method (ISORM) is applied to update the reliability index for an optimal solution. The global energy for different multi-year return periods is treated as objective functions to determine the extreme-value profiles corresponding to strength design of underwater structures. For the current model related to dynamic design, a statistical classification method based on the clustering large application algorithm (CLARA) is performed to obtain a multi-profile current model by using a large long-term prototype monitoring current dataset. An average silhouette width is found and used to determine the final classification number. Finally, a multi-profile current model is established for the underwater structural dynamic design. The presented current models can be directly applied as the current loads for underwater structural design in the South China Sea. Moreover, besides the possibility to extend the proposed method to other ocean areas, it can also be used for developing models of wind, acoustic, and other field variables.