Vector ARMAX modeling approach in multi-input modal analysis

✍ Scribed by Byeong-Ho Park; Kwang-Joon Kim

• Publisher: Elsevier Science
• Year: 1989
• Tongue: English
• Weight: 1010 KB
• Volume: 3
• Category: Article
• ISSN: 0888-3270
• DOI: 10.1016/0888-3270(89)90044-7

✦ Synopsis

This paper analyses multivariate time series using a parametric approach for the purpose of identification of modal parameters of mechanical structures. Because of computer capacity the multi-input/multi-output (MIMO) data were treated with the multiinput/single-output (MISO) consecutively. By noting that some of the modal parameters must have global characteristics regardless of measurement locations in theory, a MIMO modeling approach is taken to analyse a set of multiple-random excitation/multiple response measurements. The MIS0 approach is also applied to the same data and the results obtained by the two methods are compared. Two specific regression models for the MIMO and MIS0 approach are derived from the vector Autoregressive Moving Average model with exogenous variables (ARMAX) and the least squares method is applied iteratively for the parameter estimation. The modal parameters are derived from the parameters of the vector ARMAX based upon the principle of impulse response invariance. The procedures are used to analyse a set of simulation data of a three degree of freedom system.