Eventos Anais de eventos
DINAME 2017
XVII International Symposium on Dynamic Problems of Mechanics
Dynamical Systems Identification with Smooth Decomposition
Submission Author:
Damien Foiny , RJ , Brazil
Co-Authors:
Damien Foiny, Gustavo Brattstroem Wagner, Rubens Sampaio, Roberta Lima
Presenter: Damien Foiny
doi://10.26678/ABCM.DINAME2017.DIN17-0192
Abstract
Smooth Decomposition (SD) is a multivariate data or statistical analysis method to find normal modes and natural frequencies in an spatial data field. The projection used for this method is made such as it keeps the maximum variance possible for the displacement vector and also as it keeps the smoothest motions along time. From this method we can get the "energy" participation in the response of each normal mode during the simulation or the experimental test which can be a relevant information to validate results concerning the identification process. This method of identification can be used for linear and nonlinear systems and uses only output data given that the excitation satisfies some properties normally met by a well chosen random excitation, as a white noise, for example. The objective of this method is to identify systems from their displacement field under ambient excitation which, in many cases, can be hard to compute or to describe. As the method is only based on the covariance matrices of the displacement field and the corresponding velocity field, it is no needed further considerations and approximations. In this point the method is a great tool for modal analysis and system identification. In this paper, the presentation of the method is firstly done which will show us how we can interpret the results of SD for different systems and then the application of SD on simulated multi-DoF damped and undamped systems is performed and discussed to understand how SD can be a great tool for modal analysis. A discussion about the quality of the excitation is also performed.
Keywords
Smooth Decomposition, System Identification, Operational Modal Analysis, Nonlinear Parameters Identification
