Eventos Anais de eventos
COBEM 2019
25th International Congress of Mechanical Engineering
STRUCTURAL HEALTH MONITORING IN MECHANICAL SYSTEMS FROM CHANGES IN THEIR VIBRATION CHARACTERISTICS BASED ON ARTIFICIAL NEURAL NETWORKS AND PCA
Submission Author:
Ricardo De Medeiros , SC , Brazil
Co-Authors:
Luísa Völtz, Eduardo Lenz Cardoso, Ricardo De Medeiros
Presenter: Ricardo De Medeiros
doi://10.26678/ABCM.COBEM2019.COB2019-1630
Abstract
Structural Health Monitoring (SHM) methods comprise early detection of modifications in a given set of characteristics of a system to predict failure. Inspired by the biological nerve system, artificial neural networks have been applied in machine learning for data classification and pattern recognition and can handle a wide variety of non-linear complex systems. This work proposes the use of both principal component analysis and feedforward artificial neural networks to analyze vibrational data to predict failure. To this end, three different mechanical systems: aluminum beams, rolling bearings, and composite plates were studied in details. For all systems, different damage scenarios are considered. The results show that the methodology can detect damages in the different mechanical systems applied with excellent levels of accuracy.
Keywords
Damage Detection, Structural Health Monitoring (SHM), Artificial neural networks (ANNs), Principal component analysis (PCA), Frequency response functions (FRFs)
