The dominant excitation forces are generally measurable during the forced vibration tests of structures unlike the ambient vibration tests. Not considering of input forces in the system identification is one of the main sources for error generation in the Operational Modal Analysis (OMA). Therefore, some non-structural dynamic characteristics obtained due to the excitations effects can be eliminated by considering the input forces. In this paper, a special modal analysis is presented in the subspace method that removes the excitation effect of the measured input forces from the test data using orthogonal decomposition and identifies the system with an optimal subspace method based on canonical correlation analysis (SSI-CCA). To evaluate the proposed method, the seismic response of the Pacoima dam and forced vibration test results of the Alamosa Canyon Bridge are used. Non-structural and noisy pole removal, and increased accuracy of the extracted modal properties, specially damping ratios, can be mentioned as one of the important results of this study. Four non-structural modes are identified using the SSI-Data method while the first two modes without any noises, the same as previous results, are extracted using the proposed method. In addition, the damping ratios of the Alamosa Bridge are obtained by Hammer test, which are not obtained in the previous investigations.

Keywords: Orthogonal Decomposition, State Space, Subspace, Hankel Matrix

Full-Text [PDF 947 kb]   (5724 Downloads)