Improvements in the HHT for the Modal Parameter Identification of Structures with Closely Spaced Modes

ABSTRACT

The Hilbert-Huang transform (HHT) is widely used to identify modal parameters of structures because of its advantages in analyzing nonstationary signals. However, the mode mixing and end effects of empirical modal decomposition (EMD) and Hilbert transform (HT) cause a significant reduction in the identification accuracy. In this paper, the traditional HHT is improved to identify the modal parameters of structures with closely spaced modes, such as arch dams. First, instead of EMD, the blind source separation (BSS) and singular spectrum analysis (SSA) are proposed to obtain the generalized intrinsic mode functions (IMFs), respectively. The BSS is applied to the stationary signal such as ambient vibration data excited by wind, traffic, and water pressure fluctuation effects; and the SSA is used to extract the IMFs from the nonstationary response, such as earthquake response excited by ground motions. Second, an automatic moving-window, which automatically selects the optimal linear segment to identify the frequency and damping ratio, is proposed to reduce the end effects of HT. Finally, the effectiveness of the improved HHT is verified using the analytical solution of a three-degree-of-freedom system and the numerical simulation of the Dagangshan dam. In addition, the natural frequencies and damping ratios of the Mauvoisin dam are identified based on the measured data using the improved HHT. The results show that the improved HHT has great advantages in practical engineering.

KEYWORDS:

• HHT
• modal parameters identification
• blind source separation
• singular spectrum analysis
• automatic moving-window

Acknowledgments

The authors express their sincerest gratitude for the support. The authors also appreciate Dr. Li-Hang Kou at CHN ENERGY for this valuable discussion in communication during the research.

Additional information

Funding

This research is financially supported by the National Natural Science Foundation of China (Nos. 51725901, 51639006, and 51579133).