Eventos Anais de eventos
COBEM 2023
27th International Congress of Mechanical Engineering
Effects of Rational-Function Approximation Coefficients on the Aeroelastic Analysis of Transonic Flutter with CFD-Based Reduced-Order Model
Submission Author:
Ana Cristina Neves Carloni , SP , Brazil
Co-Authors:
Ana Cristina Neves Carloni, João Luiz F. Azevedo
Presenter: Ana Cristina Neves Carloni
doi://10.26678/ABCM.COBEM2023.COB2023-0147
Abstract
The current work is concerned with studying the effects of rational-function approximation (RFA) coefficients of the aerodynamic transfer functions on the transonic aeroelastic stability analyses performed by reduced-order models based on computational fluid dynamics (CFD) results. The CFD calculations are based on the Euler equations, and the code uses a finite volume formulation for general unstructured grids. A centered spatial discretization with added artificial dissipation is used, and an explicit Runge-Kutta time marching method is employed. The dynamic system considered in the present work is a NACA 0012 airfoil-based typical section with two degrees of freedom in the transonic regime. Unsteady calculations are performed for mode-by-mode and simultaneous excitation approaches, the latter defined by orthogonal Walsh functions. System identification techniques are employed to allow the splitting of the aerodynamic coefficient time histories into the contribution of each individual mode to the corresponding aerodynamic transfer function. Generalized unsteady aerodynamic forces are approximated by a rational transfer function in the Laplace domain, more specifically the first and second form of the Eversman and Tewari interpolating polynomials. Given that the first form is susceptible to poorly-conditioned eigenvalue problems, the second form consistently accounts for cases where the optimized values of two or more poles of the approximation are close to one another. Two approaches are implemented in order to select the nonlinear parameters present in the rational function approximations. First, the aerodynamic lag parameters are arbitrarily chosen and kept fixed during the entire aeroelastic analysis. Although this traditional approach is clearly inconvenient due to the necessity of adjusting the nonlinear parameters to each application of interest, it is implemented in this work for comparison purposes. Second, nonlinear parameters are also selected in this work through a non-gradient Nelder-Mead optimization process. Results suggest that small variations in the aerodynamic lag states identified in the rational-function approximation considerably impact the flutter onset point identification in the aeroelastic stability analysis.
Keywords
Rational-function approximation, reduced order model, Flutter, Aeroelasticity
