STRUCTURAL MODAL IDENTIFICATION USING AN IMPROVED EMPIRICAL MODE DECOMPOSITION

摘要

Empirical mode decomposition (EMD) has shown significant promises in signal decomposition of vibration data of civil engineering structures. Owing to its self-adaptive time-frequency decomposition capability, it is widely used in system identification of both linear and nonlinear structures. Unlike EMD which uses only single sensor, multivariate EMD (MEMD) is recently explored as a modal identification tool utilizing multichannel vibration measurements. In this paper, the performance of MEMD is investigated by integrating with another powerful signal separation technique to undertake modal identification under a wide range of applications. The proposed EMD method is validated using a suite of numerical studies.