Accurate and efficient, time-linearized flutter analysis of transport aircraft

Abstract

Calculating the flutter boundary by solving the Reynolds averaged Navier Stokes (RANS) equations in the time domain is prohibitive concerning the computational costs. Motion induced, unsteady aerodynamic forces have to be computed for a range of parameter combinations, e.g. Mach number, altitude, mode shape and oscillation frequency. However, the accuracy of RANS methods is required to compute flutter boundaries in the transonic flow regime, where recompression shocks and shock-induced flow separations occur. Linearized frequency domain methods are one approach to retain the prediction quality while reducing the computational complexity significantly. The application maturity of such a method is demonstrated in this work. An extension of the method into the Laplace domain is discussed to analyze the validity of the p-k method in the flutter analysis. A new iterative scheme to solve the system of linear equations is presented. This method was implemented to improve the robustness and efficiency of the linear solver especially at transonic flow conditions. The work comprises the validation of the linearized frequency domain method for motion induced unsteady air loads in comparison to its nonlinear, time-domain counterpart for the 2d NACA 64A010 airfoil, the trapezoidal AGARD LANN wing and the generic transport aircraft FERMAT.After demonstrating the capability to predict unsteady air loads accurately as well as efficiently, the method was applied to a flutter analysis of the FERMAT configuration. Two different flutter cases are found: the first is a combination of the wing and horizontal tail plane whereas the second flutter case is dominated by the horizontal tail plane. The second case is more critical in the analyzed range of Mach numbers. Moreover, the introduced error of the frozen eddy viscosity approach on the unsteady, aerodynamic loads and on the flutter results is analyzed.