Técnicas de inteligência artificial aplicadas ao método de monitoramento de integridade estrutural baseado na impedância eletromecânica para monitoramento de danos em estruturas aeronáuticas

• Ano de defesa: 2012
• Autor(a) principal: Palomino, Lizeth Vargas
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; BR; Programa de Pós-graduação em Engenharia Mecânica; Engenharias; UFU
• Palavras-chave em Português: Monitoramento de integridade estrutural; Impedância eletromecânica; Redes neurais; Análise de cluster; Condições ambientais; Sistema de medição de impedância; Aeronaves; Localização de falhas (Engenharia); Redes neurais (Computação); Structural health monitoring; Electromechanical impedance; Artificial neural network; Fuzzy cluster analysis; Environment conditions; Impedance measurement system; CNPQ::ENGENHARIAS::ENGENHARIA MECANICA
• Link de acesso: https://repositorio.ufu.br/handle/123456789/14726; https://doi.org/10.14393/ufu.te.2012.57

The basic concept of impedance-based structure health monitoring is measuring the variation of the electromechanical impedance of the structure as caused by the presence of damage by using patches of piezoelectric material bonded on the surface of the structure (or embedded into). The measured electrical impedance of the PZT patch is directly related to the mechanical impedance of the structure. That is why the presence of damage can be detected by monitoring the variation of the impedance signal. In order to quantify damage, a metric is specially defined, which allows to assign a characteristic scalar value to the fault. This study initially evaluates the influence of environmental conditions in the impedance measurement, such as temperature, magnetic fields and ionic environment. The results show that the magnetic field does not influence the impedance measurement and that the ionic environment influences the results. However, when the sensor is shielded, the effect of the ionic environment is significantly reduced. The influence of the sensor geometry has also been studied. It has been established that the shape of the PZT patch (rectangular or circular) has no influence on the impedance measurement. However, the position of the sensor is an important issue to correctly detect damage. This work presents the development of a low-cost portable system for impedance measuring to automatically measure and store data from 16 PZT patches, without human intervention. One fundamental aspect in the context of this work is to characterize the damage type from the various impedance signals collected. In this sense, the techniques of artificial intelligence known as neural networks and fuzzy cluster analysis were tested for classifying damage of aircraft structures, obtaining satisfactory results. One last contribution of the present work is the study of the performance of the electromechanical impedance-based structural health monitoring technique to detect damage in structures under dynamic loading. Encouraging results were obtained for this aim.