Reduced Order Modeling for Hypersonic Aeroelasticity
Abstract
The time required to perform dynamic simulation of realistic, complex, high-dimensional models is a bottleneck towards efficient design and optimization. We explore a few themes in model reduction with special attention to structural dynamics applications featuring geometric nonlinearities. This work promotes the use of nonlinear manifolds and simulation-free approaches for model reduction. The main contributions of this work include a method for obtaining reduced order models (ROM) of geometrically nonlinear structures via a quadratic mapping. The techniques allow for nonlinear model reduction in a mathematically justifiable manner and result in a single-degree-of-freedom yet exact ROM for the beam. Further contributions in this work pertain to hyper-reduction, essential to achieving computational speed. We propose a nonlinear extension to the energy conserving sampling and weighing method. We also propose a heuristic method that avoids the need for full simulations in the training stage. We include the effects of temperature in the nonlinear structural dynamics equations, and propose reduction via the method of multiple scales, where the reduction basis efficiently adapts to the instantaneous temperature configuration of the structure.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 27, 2019
- Accession Number
- AD1176800
Entities
People
- Paolo Tiso
- Shobhit Jain