Structural damage localization through multi-channel empirical mode decomposition

Abstract

Purpose

The purpose of this paper is to develop a new structural damage detection technique based on multi-channel empirical mode decomposition (MEMD) of vibrational response data.

Design/methodology/approach

Empirical mode decomposition (EMD) is an empirical data-based signal decomposition method which has been applied in many engineering problems. Utilizing classical EMD to reveal the damage-indicating features of structural vibration response encounters some difficulties due to the inconsistency of modes obtained from different data channels. To overcome this problem, MEMD has been employed. To this end, MEMD algorithm has been adopted to impulse response vector of measured DOFs. The proposed method has been carried out concerning both numerical and experimental beam models. Damage has been modeled by reducing the flexural rigidity in some predefined beam sections. The effects of various factors such as measurement grid density, damage severity and damage position are investigated.

Findings

The results of both numerical and experimental case studies have been promising. The method could determine the damage location in all cases. The efficiency of method gets better when damage is located far from inflation points of the corresponding mode. In such cases, utilizing higher modes can make up the efficiency.

Research limitations/implications

Since the present research is the first investigation of MEMD in damage localization, just one-dimensional structures have been studied. Extending the method to more complicated geometries needs further attempt.

Originality/value

Although a number of relevant studies have been carried out based on EMD, up to the author’s best knowledge, this is the first attempt to structural damage localization using MEMD.

Citation

Salehi, M. and Azami, M. (2019), "Structural damage localization through multi-channel empirical mode decomposition", International Journal of Structural Integrity, Vol. 10 No. 1, pp. 102-117. https://doi.org/10.1108/IJSI-07-2018-0037