Title:
A Sliding-Window Matrix Pencil Method for Aeroelastic Design Optimization with Limit-Cycle Oscillation Constraints
A Sliding-Window Matrix Pencil Method for Aeroelastic Design Optimization with Limit-Cycle Oscillation Constraints
Author(s)
Golla, Tarun
Advisor(s)
Riso, Cristina
Kennedy, Graeme J.
James, Kai
Kennedy, Graeme J.
James, Kai
Editor(s)
Collections
Supplementary to
Permanent Link
Abstract
This paper presents a new approach for constraining limit-cycle oscillations in aeroelastic design optimization. The approach builds on a gradient-oriented limit-cycle oscillation constraint that bounds the recovery rate to equilibrium, bypassing the need for bifurcation diagrams. Previous work demonstrated the constraint using recovery rates approximated via a conservative approach. This work introduces a new approach to accurately evaluate recovery rates from transient simulations. The approach uses the matrix pencil method within a time window that slides along the time history for the quantity of interest, allowing this damping identification method to resolve amplitude-variant nonlinear effects. The new sliding-window matrix pencil method is verified with reference recovery rates from envelope finite differencing of the dynamic responses induced with a large initial perturbation of a typical aeroelastic section. Sensitivity analyses identify optimal parameters to obtain accurate recovery rates while minimizing computational costs. The new developments are then demonstrated by optimizing the typical section subject to the proposed limit-cycle oscillation constraint along with flutter and side constraints. The results are compared with previous work that solved the same optimization problem by evaluating the limit-cycle oscillation constraint using approximate recovery rates. The limit-cycle oscillation constraint based on the new sliding-window matrix pencil method allows the optimizer to achieve a less conservative design solution while satisfying the constraints. This methodology was additionally extended through the optimization of a more complex 3-variable optimization. The implementation was further ported into a modular framework within which results were verified, allowing for future extensions to this methodology. This work is anticipated to pave the way for larger-scale aeroelastic design optimizations subject to limit-cycle oscillation constraints.
Sponsor
Date Issued
2023-12-15
Extent
Resource Type
Text
Resource Subtype
Thesis