A novel method of parameter identification based on nonlinear empirical model for vortex-induced vibration

The paper deals with the identification of the parameters of Scanlan's nonlinear model for the vortex-induced vibrations (VIV) of bridge decks. Although genetic algorithm (GA) has a good search ability, it has difficulties in optimizing the multi-parameter problem to obtain acceptable results. Levenberg-Marquardt algorithm (LMA), a widely used optimization algorithm, is strongly influenced by the initial condition. A novel identification method (GALMA) based on GA and LMA has been proposed. GA is adopted to achieve an initial condition, which is successively used by LMA to optimize the identification. In order to verify the validity of the identified parameters, the reconstructed data is compared with the measured data with respect to the time history of vortex-induced vertical force (VIVF), the frequency components of VIVF, and the time history of displacement response. The result shows that GALMA is effective. Furthermore, compared with other identified methods, GALMA is more precise in identification parameters, especially many parameters of equation.