Esmaeil Sabzali H., Montazeri A., Poshtan J., Jahed Motlagh M.

The aim of this paper is robust identification of a lightly damped flexible beam with parametric and non-parametric uncertainties in order to be utilized by robust control techniques such as . Fundamentally, analytic models of lightly damped flexible structures such as beam are infinite dimensional. In order to design a controller for this structure one has to have a finite dimensional model. This model truncation often leads to poor performance and stability. To overcome this problem, robust controllers are introduced. Therefore specific identification methods are necessary when one seeks to use robust control schemes. Identification methods producing a nominal model and its associated uncertainty are known as 'Robust Identification' or 'Control-oriented Identification' methods. In particular, our approach is based on worst case estimation theory where uncertainties are assumed to be unknown but bounded (UBB). This type of uncertainty representation is greatly adopted by various robust control methods. We examine and compare different outbounding algorithms (parallelotopic and ellipsoidal) for estimation of the feasible parameter set that has been delivered by the so-called set membership identification algorithm. Because of high magnitude of non-parametric uncertainties, which are high frequency unmodelled dynamics in our case, the resulting uncertainty bounds are not satisfactory. In order to proper handling with the high magnitude non-parametric uncertainties, it is shown that the combination of set membership approach with other identification techniques such as 'Model Error Modeling' will result in superior results.