Eventos Anais de eventos
COBEM 2017
24th ABCM International Congress of Mechanical Engineering
DAMAGE DETECTION IN A CANTILEVER BEAM BY A STOCHASTIC OPTIMIZATION ALGORITHM THROUGH THE VARIATION OF THE NATURAL FREQUENCIES
Submission Author:
GUSTAVO ZENI , RS
Co-Authors:
Letícia Miguel, Leandro Fadel Miguel, GUSTAVO ZENI
Presenter: GUSTAVO ZENI
doi://10.26678/ABCM.COBEM2017.COB17-0157
Abstract
This paper presents the damage detection as an optimization problem using a stochastic optimization algorithm known as Backtracking Search Algorithm (BSA). This one uses the differences from the first five natural frequencies of a cantilever beam identified from the Stochastic Subspace Identification (SSI) as parameters for the objective function on the optimization process. The beam was modeled with ten Euler-Bernoulli elements through a FE model developed by the authors and the response for free vibration and ambient vibration was recorded using the Newmark method. In order to test the accuracy and efficiency of the method, four different damage scenarios were proposed and noise of 3% and 5% was introduced on the measurements. The proposed method for damage detection was accurate for free vibration test and low level of noise in the ambient vibration. As the noise level was increased, the identification of natural frequencies was less precise, thus the damage localization indicated some elements at the free end of the beam that in fact was not damaged.
Keywords
Structural Health Monitoring (SHM), Stochastic Subspace Identification (SSI), Stochastic optimization algorithm, Backtracking Search Algorithm (BSA), Ambient vibration
