A Hilbert Transform Method for System Identification of Time-Varying Building Structures

This paper presents a Hilbert transform method for the time-varying property identification of high-rise buildings with limited sensor deployments in combination with an observer technique and a branch-and-bound technique. The observer technique is introduced to estimate building responses from the limited measurements. This estimation depends upon the current/damaged condition of buildings to be identified. As such, the branch-and-bound technique is used to develop potential damage scenarios and pinpoint the specific damage scenario that minimizes a global error index of stiffness and damping coefficients while limiting the local error of each story into less than a predetermined threshold. A 60-story shear building with abruptly varying stiffness at multiple stories is used as an example. Numerical results indicate that the proposed method can accurately detect locations and variations of the structural properties of the building with limited sensors deployed at appropriate locations. Additional research indicated insignificant effects of noise and successful detections of damage at adjacent stories