Simultaneous identification of structural time-varying physical parameters and unknown excitations using partial measurements

Abstract

Structural systems often exhibit time-varying dynamic characteristics during their service life due to severe hazards and/or environmental erosion. Therefore, the identification of time-varying structural systems is important. So far, methods based on wavelet multiresolution (WM) analysis have been proposed for the identification of structural time-varying physical parameters. However, full information on both complete structural responses and external excitations were requested in previous approaches, which greatly restricts their applications in practice. To overcome this limitation, an algorithm is proposed in this paper for simultaneous identification of structural time-varying physical parameters and unknown external excitations using only partially measured structural responses. The proposed algorithm is based on the integration of WM analysis and the Kalman filter with unknown input (KF-UI) approach recently developed by the authors. Firstly, structural time-varying physical parameters are decomposed by WM expansion, transforming the identification task into scale coefficients estimation. Then, the KF-UI approach is used for simultaneous identification of structural state and unknown excitations using partially measured structural responses. Finally, the scale coefficients are estimated by nonlinear least-squares optimization and the original time-varying physical parameters are re-constructed. Numerical simulations and an experimental test are conducted to validate the proposed algorithm.