Computational fluid dynamics based aeroelastic system identification and flutter prediction

The paper presents a computational methodology for CFD-based flutter analysis that is based on system identification of the aeroelastic system at few sub-critical dynamic pressures. Autoregressive Exogenous (ARX) models of the aeroelastic system are estimated based on simulated aeroelastic modal responses to prescribed excitation of time-varying vertical velocities. A linear stability parameter of the system is computed for each dynamic pressure, and, by extrapolation, points to the flutter conditions. The method was demonstrated on three test cases of subsonic airfoil flutter (2 DOF plunge and pitch aeroelastic system), transonic airfoil flutter (2 DOF plunge and pitch aeroelastic system), and a generic transport aircraft at subsonic flow, with flutter mechanism involving three structural modes. In all cases the method resulted in accurate predication of the flutter point as compared to either full CFD aeroelastic simulation, or wind tunnel test results. The method is computationally efficient, as it only requires a single aeroelastic simulation, at two to three dynamic pressure values, to predict the flutter point (for each Mach number). The paper describes the methodology, and discusses the details of its application, its advantages, and shortcomings.