Experiments and Simulations to Facilitate Data-driven and Machine-learning Hydrodynamic Models
Abstract
Mirjam Furth, Texas A&M University, Funds Requested: $1,000,000, Approved for Public ReleaseHigh Speed Craft (HSC) is a vessel class characterized by its ability to reach high speeds through planing. They are widely used within the Navy, from small boats without a permanent crew to the Freedom-Class Littoral Combat Ships. The performance of HSC is determined by their resistance, running position, acceleration in waves, ability to handle hull slamming pressure, maneuverability, and operability. These are governed by hydrodynamic phenomena that are not trivial to predict due to their dynamic nature.With the progression of data science and analytics, Machine Learning (ML) and Artificial Intelligence (AI) are becoming essential elements in the design and operation of naval craft, paving the way for innovations in marine vehicle design and operation. This proposal is filed jointly with related proposals from collaborators at the Naval Surface Warfare Center Newport (NUWCDIVNPT) and the University of Massachusetts Dartmouth (UmassD).The collective goal is to develop a data-driven approach for building a reduced-order predictive hydrodynamic ML model for HSC in normal operating conditions and extreme edge cases. The ML model will be developed by Dr. Hixenbaugh at NUWCDIVNPT and Dr. Heryudono at UmassD. The success of the development of this machine learning technique for predicting the performance of High-Speed Crafts relies on accessto high quality accurate data. PI Furth will provide the numerical and experimental datasets required to develop and train a physics-informed ML model. The numerical datasets will be generated by a Boundary Element Method (BEM) and the experimental datasets will be generated by free-running model tests in situ. Targeted novel simulations and experiments will be performed to obtain the hydrodynamic and forcing coefficients needed to solve a 6-Degree of Freedom (6-DoF) Fossen-type equation. PI Furth has previously developedan 8 ft free running scaled model of the Generic Prismatic Planing Hull (GPPH) which is an ideal experimental platform for large scale data generation of HSC hydrodynamic performance. The free running model is self-propelled and equipped with a data acquisition system (DAQ) to gather hydrodynamic information. The in-situ settings provide a more realistic testing environment which is a diversemix of current, wave, and wind conditions. An environmental monitoring system comprised of four commercial wave buoys and a stereovision system can be employed to measure the sea state. Anticipated outcomes include:* A step towards digitalization of the design spiral and real-time operational feedback for vessel operators through the ML model. * New and enhanced methodologies for experimentaldata collection for ocean free-running models.* A centralized data repository to store both experimental and numerical data relatedto the performance of HSC, which will serve as a valuable resource for both current and future research.These outcomes will significantly enhance our understanding of the hydrodynamic performance of HSC in various sea conditions during advanced and common maneuvers. The data and models developed through this project will facilitate a digitalization shift in ship design and operation, and ultimately enhance the safety and operability of fast planing vessels.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 09, 2024
- Source ID
- N000142412523
Entities
People
- Mirjam Furth
Organizations
- Office of Naval Research
- Texas Engineering Experiment Station
- United States Navy