Damage detection of high-rise buildings using an eigenvalue problem-based inverse solution

Abstract

Structural health monitoring (SHM) has been applied in the regular control of high-rise buildings' health that has deteriorated having being subjected to a sudden loading. Storey-level damage detection has been a subject of focus, due to the complexity of high-rise buildings. In this study, that of two-dimensional (2D) high-rise buildings is the objective of this study. The eigenvalue problem-based inverse solution is a promising method to identify the changes in the mechanical matrices of a building, once the issues related to the huge number of degrees of freedom (DOFs) can be dealt with. The Guyan static condensation procedure is applied to reduce the full matrices based on the limited size of eigenvectors measured in field. The modal data is obtained from a simple sensor network in which requires only one uniaxial accelerometer per floor. Two techniques, particularly damage detection and mass recognition, are developed, based on the inverse solution. The proposed approach is validated numerically on 20-storey and 30-storey buildings. Reliable storey-level detection is achieved as long as the modal data is noise-free or low-level noise-contaminated. Furthermore, the mass recognition procedure is successfully verified using an experimental test on a 3-storey frame.