Time-varying system identification using a newly improved HHT algorithm

Abstract

A newly improved Hilbert-Huang transform (HHT) algorithm is presented for identification of time-varying systems and analysis of nonlinear structural response with closely spaced modes. In this improved HHT, the auto-correlation function of the structural response is taken as a substitute of input to reduce noise effect. A band-pass filter scheme and an effective intrinsic mode function (IMF) selection principle are combined to overcome the modal perturbation problem. Based on these, a time-varying system identification procedure is developed. Its robustness and effectiveness are verified by both numerically simulated and laboratory measured vibration data on a scaled concrete-steel composite beam model.