Session: 02-06-01: Structural Dynamics and Control of Aerospace Structures

Paper Number: 137946

137946 - A Consistent and Conservative Modular Aeroelastic Coupling Methodology for Multidisciplinary Design Optimization 

Multidisciplinary Design Optimization (MDO) is becoming an increasingly relevant field in the search for more efficient aircraft. Since the start of the century several advances in computational mathematics and computer science have made gradient-based optimization approaches feasible. This has made it possible to solve industrial-scale MDO problems with dozens or more design variables. As was already mentioned in NASA's CFD Vision 2030, implementing multiple coupled models for use in gradient-based MDO has often required significant software development efforts. Furthermore, one-off model solver coupling approaches are often used due to the robustness and stability requirements posed by MDO problems. This has the added problem of making it difficult to generalize such model coupling approaches.

We propose a methodology for aeroelastic coupling that is modular through the use of a Solver-Independent Field Representation (SIFR), which acts as intermediate representation between fluid and solid models. This results in an indirect fluid-solid coupling approach. The exchange of information between SIFR and physics models is standardized in order to facilitate low-effort solver integration for coupled aeroelastic analysis. Moreover, the coupling methodology allows for the specific construction of coupling operators that retain aerodynamic force consistency and aeroelastic work conservation. Retaining these invariants is intended to aid in the robustness of coupled aeroelastic simulations. We derive algebraic conditions and matrix shape constraints that must be met to achieve the conservation of these invariants. We introduce ‘invariant matrices’ for computing the aeroelastic work in SIFR and each model; this allows us to implement the coupling methodology in a general yet discretization-cognizant way. The modularity of the coupling approach has the additional benefit of being well-suited for multi-fidelity optimization approaches: Either fluid or solid solver can be changed without affecting the flow of information between SIFR and the other solver.

The coupling methodology is applied to various test cases and compared to experimental data and other simulation tools. We compare the aeroelastic predictions obtained with and without the enforcement of aeroelastic work conservation.

Presenting Author: Sebastiaan Van Schie University of California San Diego

Presenting Author Biography: Sebastiaan is currently a third-year PhD student at UC San Diego in the lab of Professor John Hwang. Before starting at UC San Diego he got a bachelor and a master in aerospace engineering and a master in applied mathematics from TU Delft in the Netherlands. Prof. Hwang's lab focuses on Multidisciplinary Design Optimization (MDO). Within the scope of MDO Sebastiaan researches Reduced-Order Modeling and aeroelasticity. His other research interests include computational physics and numerical discretization techniques.