Large Scale Optimization Via Reduced Sub-Space Multipoint Approximations and Continuous Sensitivity
Abstract
The objective of this research was to derive, demonstrate, and apply new computational algorithms for the aeroelastic design of coupled structural and aerodynamic models of aircraft. An approximate response surface was derived to accelerate convergence to an optimal design using a trust-region-based sequential multipoint approximation algorithm (MCA). It was shown reduce the number of numerical simulations needed to solve nonlinear optimization problems that entail for a large number of design variables. Continuous sensitivity equations (CSE) for fluid-structure interaction (FSI) were also developed. The CSE and their associated sensitivity boundary conditions were derived for a built-up joined beam structure under transient aerodynamic loads. The local formulation was compared to the total form CSE in terms of derivation, implementation, and results. For built-up structures with stress discontinuity at the joints, the total derivative (domain velocity) formulation of CSE was shown to be easier to implement than the local derivative (boundary velocity) formulation. Finally, the coupled fluid- structure physics and CSE for gust response of a nonlinear joined beam with an airfoil model was posed and solved, using both local and total derivative formulations. The latter was shown to be more accurate. Accuracy of the results was verified by the finite difference method.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 15, 2011
- Accession Number
- ADA564483
Entities
People
- Robert A Canfield
Organizations
- Virginia Tech