This paper presents a Hilbert-Huang transform based signal reconstruction technique for the modal analysis of structural systems using vibration measurements. The original measured signal is initially undergone a well defined band-pass filter in order to solve the mode confounding problem. After the data preprocessing, each mode of the signal is reconstructed via the proper selection of intrinsic mode functions (IMFs) that are derived from the empirical mode decomposition of the signal’s mode. Through the signal reconstruction mode by mode, the structural parameter such as natural frequency is accurately evaluated, whose accuracy depends on the criterion for selecting the IMFs using the developed component sifting process. Reliable evaluation of systems’ characteristics leads to accurate prediction of systems’ behaviors for structural safety purpose. In this study, data preprocessing is operated to alleviate the problems of mode mixing and noise contaminated signal, as well as to compare with the previous work.

This content is only available via PDF.
You do not currently have access to this content.