On the identification of indicial functions from measured aeroelastic derivatives

Two wind-load models for self-excited forces on bridge decks are compared: (1) the mixed frequency-time domain model via aeroelastic derivatives, and (2) the pure time-domain model via indicial functions. The theoretical equivalence of the two approaches is demonstrated and computational advantages of indicial functions are pointed out. Indicial functions are analytically approximated by a sum of exponential filters, whose coefficients can be related to aeroelastic derivatives and identified by means of nonlinear optimization techniques. Different algorithms are tested in order to define a robust numerical identification procedure. Some examples are presented and verified by comparison with experimental data, proving the feasibility of the approach. According to this first study, identified indicial functions can be confidently used as long as they lead to good approximation of aeroelastic derivatives. Though further investigations are required, different sets of coefficients, which well approximate measured derivatives, seem in fact to produce also good results in time-domain simulations.