A non-iterative structural damage identification methodology using state space eigenstructure assignment. Issue 2 (1st February 2017)

• Record Type: Journal Article
• Authors: Sen, Subhamoy; Bhattacharya, Baidurya
• Is Part Of: Structure and infrastructure engineering. Volume 13:Issue 2(2017)
• Page Start: 211
• Page End: 222
• Publication Date: 2017-02-01
• Subjects: Damage identification -- state space modelling -- subspace identification -- eigenstructure assignment -- structural health monitoring; Structural analysis (Engineering) -- Periodicals; Structural engineering -- Periodicals; Buildings -- Performance -- Periodicals; 620.005
• DOI: 10.1080/15732479.2016.1157825
• Languages: English
• ISSNs: 1573-2479

Abstract

Abstract : In this article, a novel approach to damage identification (location as well as intensity) is presented using eigenstructure assignment (ESA)-based finite-element model (FEM) updating. ESA is a control-based approach that utilises state or output feedback of a system to alter its eigenstructure. The proposed method identifies the system's state transition matrix and its eigenstructure from the response time history. The identified eigenstructure is first mapped onto the physical space and then reconstructed in state space in the preferred orientation and order which, in turn, is used as the target for the ESA algorithm to uniquely update the system matrices of the baseline FEM. Comparing the updated stiffness matrix with the baseline, the location and intensity of damage are estimated. Numerical validation of the method is performed on a shear frame, a Euler–Bernoulli beam, and an aluminium plate. A parametric study involving different levels of noise in the simulated response histories is undertaken. The algorithm is then tested with actual response histories from a damaged (notched) two span continuous steel beam and a damaged (indented) aluminium plate in the laboratory. The accuracy of the method in identifying the location and extent of damage is found satisfactory. Being eigenstructure based, the proposed methodology is restricted to linear time-invariant systems.