Published: Jan 30, 2023 by Daning Huang
Abstract: The multi-disciplinary analysis and optimization is becoming indispensable for modern high-performance engineering systems. However, this class of problem usually comes with an extremely high computational cost, even for the analysis of just one configuration; let alone a systematic parametric study for optimization. Reduced-order models (ROMs) are typical approaches to reducing the computational cost to a tractable level. However, the existing ROMs suffer from the curse of dimensionality that roots from the need to parameterize and sample the system configuration parameters. This work presents our recently developed methodology of physics-infused ROM (PIROM), where a first-principle-based analytical component is augmented with a data-driven physically-interpretable component. Both components admit an algebraic-differential form for enhanced modeling capability for nonlinearity. A demonstration is presented for hypersonic aerothermal load prediction, a key ingredient for the aerothermoelastic analysis and optimization of hypersonic vehicles. The PIROM achieves an accuracy close to the high-fidelity solver with excellent sampling efficiency and extrapolation capability. These results open up a new approach for the analysis and optimization of multi-physical systems.
The presentation is based on our recent work AIAAJ2023; related page here.