Reduced Order Modeling of Two-Way Ship-Airwake/Aircraft Interactions
Abstract
Efficient real-time simulation models for aircraft dynamics in ship-airwake environments are long sought after for the purposes of design, pilot training, and development of robust control. The current state-of-the-art enables near real-time one-way flight dynamics simulations in existing ship-airwakes, as well as expensive high-fidelity simulation of two-way interactions between a ship-airwake and vehicle maneuvering within it. However, reduced order modeling methods are currently inadequate for efficient simulation of the two-way interaction problem. This research is focused on this problem through novel combination of multi-scale methods from computational mechanics with machine learning. The key innovation is a hierarchical global-local decomposition approach that will substantially reduce the training requirements of a model while also increasing its generality and efficiency. This is a result of transforming the learning task from a broadly scoped problem with narrowly scoped training data - where machine learning methods struggle - to a narrowly scoped problem with broadly scoped training data - where machine learning methods excel. The hierarchical global-localdecomposition achieves this through exploitation of spatiotemporal redundancies that exist at a local level; a fundamental propertyof complex behavior in natural systems. Overall, this will enable robust and efficient simulation of two-way ship-airwake/aircraft interactions with low data requirements for model training. It will also enable incorporation of additive complexities, such as shippitching and rolling, as well as fusion of data streams from simulations, experiments and in-field measurements. This will lead to:superior physics-based flight simulators, important insights to design teams at a significantly reduced cost compared to experiments and high-fidelity simulations, and improved control design/evaluation of US Naval aircraft.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 15, 2023
- Source ID
- N000142312398
Entities
People
- Jack J. McNamara
Organizations
- Office of Naval Research
- Ohio State University
- United States Navy