An improved time-frequency analysis method for structural instantaneous frequency identification based on generalized S-transform and synchroextracting transform

Abstract

In this paper, an improved time-frequency analysis (TFA) method based on generalized S-transform (GST) and synchroextracting transform (SET), named improved synchroextracting generalized S-transform (ISEGST) is proposed to improve the accuracy of structural instantaneous frequency (IF) identification. Namely improved generalized S-transform (IGST), a new transform is first put forwarded by optimizing the window function of GST. The parameters of window function in IGST are obtained by the energy concentration measure (CM) and the optimization methodology. Applying IGST in association with SET, the ISEGST is then derived. Compared with the TFA methods in the literature, the performance of ISEGST in numerical signals has been significantly improved, especially in the presence of noise. In simulation, a nonlinear Duffing system and a two-layer shear frame model are used respectively to verify the effectiveness of the proposed method. Moreover, a time-varying cable test is conducted. Acceleration responses of the structure under linear and sinusoidal varying tension force are collected respectively. Then, the ISEGST is used for IF identification. Numerical simulation and experimental results show that the ISEGST can effectively identify IF of structures with non-stationary response signals, and it has high accuracy and good stability.