Model updating based on nonlinear normal modes identified under broadband excitation

The present paper investigates the updating of numerical models based on nonlinear normal modes (NNMs) extracted using phase separation. The proposed model updating procedure comprises three steps. First, a broadband excitation signal is applied to the structure of interest and input-output data are collected. Second, NNMs are identified by integrating a nonlinear subspace identification method and numerical continuation in a phase separation approach. Third, model parameters are updated by minimizing the difference between numerically-predicted and experimentally-estimated NNMs calculated at multiple energy levels. A numerical cantilever beam with geometrical nonlinearity is exploited herein for demonstration purposes. Synthetic vibration data are generated under a white-noise excitation. The performance of the model updating procedure is verified versus NNM identification issues, like noise perturbations.