Optimal parameters in Data-driven stochastic subspace identification in Operational Modal Analysis

Stochastic subspace identification operational modal analysis is a recently-developed method in modal testing. In this method, determining the dimensions of the Hankel matrix in the projection procedure is an important step. In this paper, the number of rows of the Hankel matrix is recommended to be larger than a half period of Toeplitz matrix. There is no need to adopt excessive number of rows in the Hankel matrix, because the projection results would not vary significantly under high number of the matrix rows. Also, if computational power is limited, the projection results stabilize more quickly if the number of rows, number of columns and the sampling frequency are decreased as the same time.