The Procedure for Subspace Identification Optimal Parameters Selection in Application to the Turbine Blade Modal Analysis

This paper considers a problem of parameters selection in stochastic subspace identification algorithm (SSI). There are two user-defined parameters in SSI, such as the “window length” and the model order. To introduce the model order selection, the stabilization diagram (SD) in combination with the modal similarity index and stabilization criteria is used. Furthermore, SD is adopted to indicate spurious modes resulting from noise. Some singular spectral analysis features are used for the optimal window length selection. The algorithm has been tested with the data obtained from the turbine blade modal testing. The obtained modal parameters in combination with blade FE-model are used to simulate the dynamic response due to multiharmonic loading. Good correlation between simulation and corresponding experimental results confirms the correctness of the proposed algorithm.