摘要
Hybrid damping mechanisms, which combine frictional and viscous damping, significantly influence the dynamic behavior of mechanical systems. However, accurately modeling hybrid damping is challenging due to its nonlinear and complex nature. To address this challenge, a recent study integrated sparse identification of nonlinear dynamics (SINDy) with recursive sine functions (ReSiF) to enable the simultaneous identification of frictional and viscous damping, as well as system stiffness, using transient response data from free-decay experiments. Unlike traditional nonlinear system identification methods, SINDy-ReSiF does not require a stable excitation source, allowing it to be used in experimental conditions where consistent excitation is difficult to maintain. Despite its potential for structural system identification, the sensitivity of SINDy-ReSiF to measurement noise, numerical differentiation errors, and initial excitation conditions has not been investigated. This study conducts numerical experiments on a two-degree-of-freedom mechanical slider to assess the impact of these factors on identification accuracy. In the experiments, response data are generated by numerically integrating the exact equations of motion. SINDy-ReSiF is then applied to the data to inversely identify the system parameters, which are subsequently compared with their exact values under various parametric conditions. The findings enhance the understanding of SINDy-ReSiF's robustness and provide a foundation for future experimental validation.