Automated Modal Analysis Using Stochastic Subspace Identification and Field Monitoring Data

The accurate identification of modal parameters is essential for structural health monitoring (SHM), as it provides critical insights into the presence of damage or degradation within the structure. A promising technique, stochastic subspace identification (SSI) has numerous advantages in operational modal analysis (OMA), particularly in implementing automated OMA. Hence, an improved procedure is proposed in this study, addressing the size of the SSI matrix, the estimation of system order, and the removal of spurious modes for automated modal analysis. A general instruction for user-defined parameters is first reviewed and summarized. Subsequently, a proposed procedure is then introduced and framed into three steps. Key advances include the preliminary identification of fundamental frequency, which helps the overall automated work, adequately assigning the size of the SSI matrix, which can improve decomposition, and a decay function, which provides a good estimation of system order. To demonstrate and verify the procedure, a numerical simulation of a ten-story shear-type building structure and two field datasets, collected from reinforced concrete (RC) frames in Taiwan, are utilized. Consequently, the results suggest that the proposed three-step procedure based on SSI can facilitate automated OMA for continuous and long-term SHM, in terms of autonomously adjusting user-defined parameters.