Abstract
This paper presents a new experimental method for modelling of a single (one)-link flexible arm in which the system identification and model reduction are combined. A linear model of the flexible arm is first derived by least-squares estimation. After identifying a model whose order is sufficiently high, the model is reduced by balanced realization. The obtained eighth-order models can represent the characteristics of the flexible arm below 100 Hz. Compared with modelling using modal analysis, the method presented does not need additional experimental equipment, such as an impact hammer or fast Fourier transform analyser, for obtaining data. The validity of the obtained models is confirmed by vibration control experiments using optimal control synthesis.
Get full access to this article
View all access options for this article.
