عنوان مقاله [English]
Much attention has been given to structural damage detection in recent decades in order to assess the reliability of structures during their service time. To detect damage in structures, one method, among different ones, is considered the most important i.e. the vibration-based methods. Because the modal parameters of structures like frequency and mode shape are so sensitive to structural properties like stiffness, it can therefore be used for detecting damage in structures.This paper presents a novel approach for structural damage detection and estimation using expanded mode shapes and extreme learning machine (ELM). One of the problems in damage detection is the compatibility between the number of sensors and Degree of Freedoms (DOFs) in the finite element model of structures, in which the number of sensors, installed to structure, is usually less than the number of DOFs in the finite element model. So, the model reduction method should be used to match incomplete measured mode shapes or the
measured mode shapes should be expanded to the dimension of the analytical mode shapes. In this study, the second option is adopted, using the improved reduction system (IRS) transformation matrix and used as input parameters to the ELM for damage identification. The proposed method uses the first two expanded mode shapes and natural frequencies as the input parameters and damage states as output to train the ELM model. Also, noise effect on the measured modal data has been investigated. The present method is applied to three examples consisting of a four span continuous beam, plane steel truss and four story plane frame. The obtained results demonstrated the accuracy and efficiency of the proposed method using incomplete modal data. Also, the results obtained indicate that the proposed method is a promising procedure for damage identification in spite of use of noisy modal data.