Influence of model order and number of block rows on accuracy and precision of modal parameter estimates in stochastic subspace identification

Abstract

The accurate estimation of the modal properties of civil structures in operational conditions is critical in a number of applications, including structural health monitoring. The available algorithms allow confident estimations of modal frequencies and mode shapes. Damping estimates, instead, are jeopardised by large error bounds, related to the non-linear behaviour of damping as well as to inherent limits of the estimators and eventually poor measurements. In general, the variability of modal parameter estimates is due not only to environmental factors but also to the performance of the estimator. Thus, the accuracy of the estimator has to be optimised to avoid errors in automated modal parameter identification. In this paper, the influence of model order and number of block rows on the accuracy of estimates via stochastic subspace identification (SSI) is investigated. An approach able to provide narrow error bounds and improve the reliability of SSI-based automated modal identification algorithms is discussed.