Time-domain system identification methods for aeromechanical and aircraft structural modeling

Time-domain model structures and algorithms are described that are suitable for the identification of aeromechanical and aircraft structural models from input-output data. An efficient batch subspace identification algorithm and its online version are presented. The batch algorithm was implemented and evaluated for structural mode identification of the V-22 tilt rotor and for the identification of models of aeromechanical instability. The online algorithm is demonstrated with a numerical example in which structural modes are identified and tracked as they appear. Our results and comparisons with current aircraft industry practice show several advantages of time-domain subspace methods over Prony and frequency-domain methods. Specific advantages include the ability to identify multiple structural modes simultaneously from a single experiment, the ability to use online system identification, and the ability to identify open-loop systems from closed-loop experimental data. These advantages have a significant effect on the number of ground and flight tests required and on how to perform testing in unstable flight regimes.