Eventos Anais de eventos
COBEM 2019
25th International Congress of Mechanical Engineering
DAMAGE IDENTIFICATION ON COMPOSITE BEAMS USING COMPUTATIONAL VIBRATION MODEL WITH UNCERTAINTIES AND ARTIFICIAL NEURAL NETWORKS
Submission Author:
Pedro Reis , SC
Co-Authors:
Pedro Reis, Luísa Völtz, Eduardo Lenz Cardoso, Ricardo De Medeiros
Presenter: Pedro Reis
doi://10.26678/ABCM.COBEM2019.COB2019-1437
Abstract
All engineering structures are subject to variations in their mechanical properties associated with damages, such as time of use, misuse, design errors, assembly and fabrication failures, climate change and other factors related to the environment. However, reliable tools that permit monitoring damage, calculating the residual resistance of the structure, allowing for possible failures to be foreseen have been sought for years by researchers and engineers. In this context, Structural Health Monitoring (SHM) is already highlighted with some applications around the world, through the identification of damage by non-destructive testing (NDT), among them, models based on the vibration response applied to Artificial Neural Networks (ANNs). However, as this method is quite sensitive to the type, quantity, and treatment of data inserted in the network, some applications are implemented using finite element (FE) methods, which, however, offer limitations on the uncertainties in the real model. Considering the scenario pointed above, this work proposes a methodology to evaluate uncertainties contained in a real model and to insert them into the computational one. After that, it is treated by Principal Component Analysis (PCA) to serve as the input of Neural Network, which has its topology determined via Particle Swarm Optimization (PSO). Thus, a multi-layer neural network was developed for detecting damage in composite beams made of Glass Fiber Reinforced Polymer (GFRP). Finally, it is discussed the potentialities and limitations of the methodology for use in damage detection systems.
Keywords
Artificial Neural Networks(ANNs), Composite beam, Damage Detection, Structural Health Monitoring (SHM), Vibration-based method
