Structural Health Monitoring Using Improved Subspace Identification Method by Including Rotational Degrees of Freedom

In the present study, system identification and subsequent damage detection of support-excited multi-story linear time-invariant structural systems are carried out using a data-driven subspace identification method. Traditionally, buildings are idealized as shear buildings and hence only translational degrees of freedom are considered. The effect of including rotational degrees of freedom in the identification technique is studied for improvement in accuracy of the identified stiffness matrix. The study is conducted on the Phase-I IASC-ASCE structural health monitoring benchmark problem using earthquake-simulated data. Since it is not possible to measure the rotational dynamic response of structural systems, these are considered as unmeasured degrees of freedom and the identification has been performed as a special case of the limited sensors problem. The HRZ lumping method is used to assign masses at the rotational degrees of freedom. An iterative technique has been proposed to obtain the full mode shape matrix, including rotational components, from the available response data at the translational degrees of freedom. Different damage scenarios have been considered and the extent of improvement in the accuracy of the stiffness matrices with better damage localization after considering rotational degrees of freedom has been demonstrated. To verify the results, different recorded earthquake data are used for the analysis of the simulated benchmark building. The robustness of the technique in the presence of noise has also been studied.