This paper presents a subspace system identification algorithm for time-invariant structural system. System matrices are estimated from input-output data. The main computational tools are the LQ and the singular value decomposition in subspace system identification. The accuracy and stability of estimated system matrices are influenced by the parameters of Hankel matrix. For stable SI, the optimal parameters of Hankel matrix are determined by considering the computational cost and the accuracy of system matrix.

• 박승근(정회원, 동아대학교 건설기술연구소 연구전담교수) | Park, Seung-Keun
• 이해성(정회원, 서울대학교 건설환경공학부 교수) | Lee, Hae Sung
• 박현우(정회원, 동아대학교 토목공학과 조교수) | Park, Hyun Woo