An Enhanced Adaptive Chirp Mode Decomposition for Instantaneous Frequency Identification of Time-Varying Structures

Abstract

Adaptive chirp mode decomposition (ACMD) is a novel decomposition algorithm, which can decompose the multicomponent signals and identify its instantaneous frequencies (IFs) simultaneously. It is crucial for ACMD to predefine the decomposition parameter (i.e., the initial IF). However, the nonstationary signals with closely spaced frequency modes or heavy noise render it difficult to preset the initial IF based on the Fourier spectrum. To overcome the difficulty of predefining the initial IF of the ACMD, a tractable version of the ACMD, termed as the enhanced adaptive chirp mode decomposition (EACMD), is proposed in this study. The EACMD adopts an adaptive method based on the maximum entropy power spectrum to automatically determine the initial IF. A simulated dynamic signal is used to evaluate the characteristics and performance of the EACMD. Furthermore, the proposed EACMD is employed to identify the IFs of time-varying structures. The effectiveness of EACMD is illustrated and verified by using a 3-story shear building model and a cable with time-varying tension forces. The results show that the EACMD can overcome the difficulty of predefining the initial IF of the ACMD and is suitable for identifying the IFs of time-varying structures, providing more accurate results than the existing techniques.