Physics-informed, machine learning methods for the quantification of extreme ocean events for naval vessels

Abstract

Physics-informed, machine learning methods for the quantification of extreme ocean events for naval vessels. The scope of this project is the development and application of new methods that combine physical modeling and machine learning, for the quantification of extreme events such as displacements and loads, experienced by naval vessels. Our goal is the formulation of a digital twin for the characterization of dynamical responses and extreme events. This is a natural extension of a long term effort between the PI and research personnel at NSWCCD through the ONR summer faculty program, for the quantification of extreme waves in the ocean environment, due to nonlinear dynamical instabilities. This is a topic of utmost importance since a deep and thorough understanding of such motions is essential for defining the safe operating envelope of naval vessels. Out effort will focus on two main tasks: I) Prototype models for extreme events in ship responses and loads ; and II) A machine learning model (digital twin) for prediction and statistical quantification .

Document Details

Document Type
DoD Grant Award
Publication Date
Apr 29, 2020
Source ID
N000142012366

Entities

People

  • Themistoklis Sapsis

Organizations

  • Massachusetts Institute of Technology
  • Office of Naval Research
  • United States Navy

Tags

Readers

  • Distributed Systems and Data Platform Development
  • Marine Hydrodynamics
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Neural Networks