MODAL PARAMETER IDENTIFICATION OF TIME-VARYING SYSTEMS USING THE TIMEVARYING MULTIVARIATE AUTOREGRESSIVE MODEL

摘要

A new time-varying multivariate autoregressive (TVMAR) model method for modal parameter identification of linear time varying (TV) systems with multi-output is introduced. Besides, a modified recursive least square method based on the traditional one is presented to determine the coefficient matrices of the TVMAR model. In the proposed method, multi dimensional nonstationary response signals of the vibrating system can be processed simultaneously. Not only the TV modal frequency and damping ratio of the system, but also the changing behavior of the mode shape in the course of vibration are identified by the proposed procedure. Numerical simulations, in which a three-degree-of-freedom system with TV stiffness is respectively subjected to impulse excitation and white noise excitation, are presented. The validity and accuracy of the method are demonstrated by the good simulation results. Keywords: modal parameter; time-varying system; TVMAR model; nonstationary