Damage Identification of Structures Using Second-Order Approximation of Neumann Series Expansion

Abstract

In this paper, a new method proposed for structural damage detection from limited number of sensors using extreme learning machine (ELM). One of the main challenges in structural damage identification problems is the limitation in the number of used sensors. To address this issue, an effective model reduction method has been proposed. To condense mass and stiffness matrices, the second-order approximation of Neumann series expansion (NSEMR-II) has been used. Mode shapes and frequencies of damaged structures and corresponding generated damage states used as input and output to train extreme learning machine, respectively. To show the effectiveness of presented method, three different examples consists of a truss structure, irregular frame and shear frame have been studied. The obtained results show the ability of the proposed approach in identifying and estimating different damage cases using limited numbers of installed sensors and noisy modal data.