Modal Identification of the "Steel-Quake" Structure Using the Data-Driven Stochastic Subspace and ARMAV Methods

In this paper, two techniques used for modal identification of output-only systems are presented and compared. The first technique is based on an ARMAV model. The method is known as the prediction error method (PEM) and requires a non-linear iterative optimisation procedure. The second technique is a stochastic subspace method that estimates the system matrices of a stochastic state space model by a data-driven algorithm and by using numerical techniques such as singular value and QR decompositions. The comparison between both techniques is performed over the “Steel-Quake” benchmark proposed in the framework of COST Action F3 “Structural Dynamics”. The results show that the investigated techniques give good results in term of estimated modal parameters. Especially, it is found that the stochastic subspace technique is much faster than the PEM.