Stiffness Identification of High-Rise Buildings Based on Statistical Model-Updating Approach

A system identification problem is investigated for high-rise buildings to identify the story stiffnesses of a shear-bending model (SB model). In the previously proposed stiffness identification method due to the present authors, the shear and bending stiffnesses of the SB model were identified by means of the subspace and inverse-mode methods. The lowest mode of horizontal displacements and floor rotation angles of the objective building was identified first by using measured data of both horizontal and rotational accelerations via the subspace method. Taking into account the resolution in the measurement of floor rotation angles in lower stories, floor rotation angles in all stories were predicted from the floor rotation angle at the top floor. However, it was difficult to obtain the bending stiffnesses reliably in the previous method. In this paper, to overcome the difficulty in the stiffness identification method using the SB model, a statistical model-updating approach is proposed, where the probability distribution of floor rotation angles in the lowest mode is obtained for the identified SB model, and a conditional probability problem is applied by providing additional measured data on floor rotation angle. The proposed stiffness identification method is useful for the structural health monitoring of high-rise buildings. For investigation of the validity of the proposed stiffness identification method, a 10-story plane-building frame is examined under micro-tremor.