Flutter and post-flutter constraints in aircraft design optimization

Jónsson, Eiríkur; Riso, Cristina; Lupp, Christopher A.; Cesnik, Carlos E. S.; Martins, Joaquim R. R. A.; Epureanu, Bogdan I.

doi:10.1016/j.paerosci.2019.04.001

Cited by 130 publications

(50 citation statements)

References 163 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This pseudoinverse is well defined since the columns ofV 1 are linearly independent. Note that this pseudoinverse appears on the right-hand side in the filtered case (5) whereas the inverse of A 1 appears on the left-hand-side in noise-sensitive simple method (2). It can be shown that pre-or post-multiplyingV T 2 by the pseudoinverse V T 1 + produces matrices with the same non-zero eigenspectrum.…”

Section: Minimum Damping Estimates Via Prony Seriesmentioning

confidence: 99%

“…Numerical flutter identification involves finding the dynamic pressure at which the minimum damping of an aeroelastic system is zero. While frequency-domain methods can be more computationally efficient [1], time-domain flutter identification remains an important computational tool [2,3]. In this technical note, we formulate a minimum damping estimate based on the matrix pencil method and demonstrate how to compute its derivative.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Flutter Boundary Identification from Time-Domain Simulations Using the Matrix Pencil Method

2019

View full text Add to dashboard Cite

show abstract

Section: Minimum Damping Estimates Via Prony Seriesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Flutter Boundary Identification from Time-Domain Simulations Using the Matrix Pencil Method

2019

View full text Add to dashboard Cite

show abstract

“…A critical aspect in constraint formulation is the constraint smoothness, which is paramount in gradient-based optimization. Jonsson et al [27] discuss many of the considerations that are necessary to formulate an efficient and continuous flutter constraint suited for a gradient-based optimization framework. Here a short summary of those considerations is also given.…”

Section: Flutter Constraint Formulationmentioning

confidence: 99%

Computational Modeling of Flutter Constraint for High-Fidelity Aerostructural Optimization

Jónsson

Mader

Kennedy

et al. 2019

AIAA Scitech 2019 Forum

Self Cite

View full text Add to dashboard Cite

High fidelity computational modeling and optimization of aircraft has the potential to allow engineers to produce more efficient designs, requiring fewer unforeseen design modifications late in the design process. In order for the optimization algorithm to generate a useful design, all the relevant physics must be considered, including flutter. This is especially important for the high-fidelity aerostructural optimization of commercial aircraft, which is likely to result in wing designs that are prone to flutter. To address this issue, we develop a flutter constraint formulation suitable for gradient-based optimization. This paper investigates the feasibility of using a Doublet-Lattice Method (DLM) based flutter constraint for high-fidelity aerostructural optimization. An efficient non-iterative root finding method is developed to compute the flutter eigenvalues from a generalized eigenvalue problem, based on the p-k flutter equation. A mode tracking scheme is implemented at two levels that successfully tracks the mode migration. An effective constraint curve is then developed to define the flutter free flight envelope, in addition to preventing discontinuities in the flutter constraint. This method allows for minimum flutter velocity to be specified implicitly. The Kreisselmeier-Steinhauser (KS) function is used to aggregate the difference of the flutter damping eigenvalues and the constraint curve into a single value that is then used in an optimization. We compute accurate and efficient derivatives of the constraint value with respect to structural sizing variables, as well as wing planform variables. Derivatives are computed using a combination of analytic and automatic differentiation methods in reverse (adjoint) and validated using the complex-step method. To study the behavior of the flutter constraint using this method, we perform a design space analysis and optimize an idealized wing (flat plate) of a rectangular planform.

show abstract

“…For gradient evaluation methods, we cover topics with a broader scope. For more detailed references on flutter analysis and gradient evaluation methods, we refer the reader to a recent review paper of Jonsson et al [5].…”

Section: Introductionmentioning

confidence: 99%