2019/12/30 by Kenan Šehić, Henrik Bredmose, Šehić, Kenan +5
Earth and Planetary Sciences · #Data Analysis #FOS: Physical sciences #Fluid Dynamics (physics.flu-dyn) #Ocean Waves and Remote Sensing #Oceanographic and Atmospheric Processes #Statistics and Probability (physics.data-an) #Tropical and Extratropical Cyclones Research
paper · pdf · doi:10.48550/arxiv.2001.03163
openalex publication_date 2019/12/30 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
We model shallow-water waves using a one-dimensional Korteweg-de Vries\nequation with the wave generation parameterized by random wave amplitudes for a\npredefined sea state. These wave amplitudes define the high-dimensional\nstochastic input vector for which we estimate the short-term wave crest\nexceedance probability at a reference point. For this high-dimensional and\ncomplex problem, most reliability methods fail, while Monte Carlo methods\nbecome impractical due to the slow convergence rate. Therefore, first within\noffshore applications, we employ the dimensionality reduction method called\n\Active-Subspace Analysis. This method identifies a low-dimensional\nsubspace of the input space that is most significant to the input-output\nvariability. We exploit this to efficiently train a Gaussian process that\nmodels the maximum 10-minute crest elevation at the reference point, and to\nthereby efficiently estimate the short-term wave crest exceedance probability.\nThe active low-dimensional subspace for the Korteweg-de Vries model also\nexposes the expected incident wave groups associated with extreme waves and\nloads. Our results show the advantages and the effectiveness of the\nactive-subspace analysis against the Monte Carlo implementation for offshore\napplications.\n