An invaded nonlinear system identification method was proposed to determine the concerned dynamic characteristics of complex jointed structures. Based on the response sensitivity analysis, a constrained nonlinear dynamic equation was developed to improve the identification performance of strongly nonlinear system. In this process, the identified dynamic characteristics and concerned nonlinear vibration responses were simultaneously defined as unknown variables to construct the extended iteration vector, which was used to solve the constrained nonlinear equation. The least-squares fitting method was used to match the dimensions of the identified dynamic characteristics and concerned nonlinear vibration responses, while the arc-length continuation method was also used to ensure the iteration convergence of the identification process. A complex thin-walled jointed structure with four cubic-stiffness models was numerically simulated to validate the proposed system identification method by a comparison with the direct iteration method. The proposed method indicated a better performance of identifying the external excitations of multi-branch nonlinear solutions, besides the concerned nonlinear vibration responses and identified dynamic characteristics were converged synchronously. Experimental investigations of a metal-rubber isolator system were also performed to validate the identification performance of nonlinear joint properties, which demonstrated a good performance of nonlinear dynamic model updating.
Wang et al. (Thu,) studied this question.