Identification of a Nonlinear Aeroelastic Aircraft Wing Model

Abstract

‡The “reverse path” spectral method, a frequency domain based nonlinear system identification technique, is considered for identifying nonlinear aeroelastic models. The method is based upon a multi-input/multi-output spectral conditioning process that mutually uncorrelates the linear and nonlinear components of the system response. Conditioned frequency response estimates derived from this process enable estimation of the system’s underlying linear frequency response matrix without contaminating effects from the nonlinearities. Analytic functions for describing the system’s nonlinearities are also identified once the linear frequency response matrix has been estimated. Since these analytic functions must be chosen prior to the identification, conditioned coherence functions, also derived from the spectral conditioning process, are used to indicate the presence (or absence) of select nonlinear components in the response data, hence overcoming the a priori assumption of this and many other parametric-based nonlinear system identification techniques. To illustrate the performance of the identification method, numerical data is first simulated from a nonlinear analytical aeroelastic system. Results show that the method is capable of identifying an accurate nonlinear aeroelastic model even when the underlying nonlinearities are unknown. The usefulness of the conditioned coherence functions for selecting analytic functions for describing the nonlinearities is also illustrated. The method is then applied to test data collected from the Active Aeroelastic Wing (AAW) aircraft, an experimental vehicle used to investigate active flexibility control of aircraft wings for improved maneuverability, weight reduction and extended range.