A data-driven approach for damage detection: an application to the ASCE steel benchmark structure

Abstract

Damage detection of systems represents an important research topic widely investigated. In particular, in the last years, several applications have been performed with reference to civil constructions in order to detect structural damages by monitoring the dynamic response. Thus far, the scientific literature has proposed many damage detection methodologies able to detect damage in structures on the basis of their time-dependent response to dynamic excitations. The aim of this paper is to derive a simple procedure for estimating the damage in structures based on a data-driven subspace identification technique. The approach provides an iterative procedure devoted to determine damage coefficients varying from zero to one and defining the reduction of the stiffness—and/or the damping—matrix from the undamaged to the damaged state of the system. A residual function based on the difference between the effective frequencies, deduced from a preliminary data-driven procedure, and the frequencies estimated during the iterations is introduced; the minimization of the residual function provides the values of the damage coefficients. In order to validate the proposed procedure and assess its advantages and limitations, the paper discusses some numerical applications carried out on the IASC-ASCE steel frame provided by the Earthquake Engineering Research Laboratory of the British Columbia University by considering different damage scenarios.