A Computational Framework for Full Resolution of Extreme Ship-Motion Statistics

Abstract

A Computational Framework for Full Resolution of Extreme Ship-Motion StatisticsWave-induced extreme motions are rare events for ships at sea. The largest ship motions can be caused by the impact of an extreme wave, or some normal waves with certain frequencies that excite the resonance of the ship. Due to complicated physical causal factors, extreme motions may occur unexpectedly and cause catastrophic damage to ships. The physical understanding and statistical quantification of extreme ship motions in a sea state are of vitalimportance to the survivability and functionality of ships. This is especially important for the next-generation naval vessels, which must complete missions in harsh sea states and even the arctic.With the increase of computational power and development of advanced algorithms, the dynamics of an individual extreme event can be studied with high-fidelity simulations (e.g.,computational fluid dynamics, CFD). However, the resolution of their statistics requires expensive Monte-Carlo method considering all incident waves, and this is computationally intractable due to the rareness of the events of interest. Reduced-order statistical approaches, such as extreme value theory/extrapolation, and design loads generator, rely on questionablestatistical assumptions and oversimplified dynamics. They are generally insufficient for the extreme ship motions, where the complex nonlinear dynamics (of wave field and wave-ship interaction) underlies the statistics.In this project, we propose a new computational framework that directly resolves the statistics (and causal factors) of extreme ship responses in a nonlinear wave field. The development leverages a range of physics and learning based approaches, including nonlinear wave simulations (potential flow), ship response simulations (e.g., CFD), dimension-reduction techniques, sequential sampling, Gaussian process regression (Kriging) and multi-fidelity methods. The key features of the new approach include (i) description of the stochastic wave field by a low-dimensional probabilistic parameter space, and (ii) use of minimum number of CFD simulations to provide most information for converged statistics of extreme motions. The two procedures enable an accurate computation of the probability distribution of the extreme ship motions with affordable computational cost. Due to the generality of the proposed framework, it can be applied to different problems associated with wave-induced extreme events, e.g., extreme loadings and green water problems.

Document Details

Document Type
DoD Grant Award
Publication Date
Feb 17, 2020
Source ID
N000142012096

Entities

People

  • Yulin Pan

Organizations

  • Board of Regents of the University of Michigan
  • Office of Naval Research
  • United States Navy

Tags

Readers

  • Computational Fluid Dynamics (CFD)
  • Distributed Systems and Data Platform Development
  • Marine Hydrodynamics

Technology Areas

  • Space