Detecção e localização de danos em materiais compósitos aplicado em aeronaves utilizando redes neurais artificiais

• Ano de defesa: 2014
• Autor(a) principal: França, Altair de Araujo [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Materiais compostos; Lamb, Ondas de; Redes neurais (Computação); Piezoeletricidade; Composite materials
• Link de acesso: http://hdl.handle.net/11449/111083

The increasing use of composite materials has brought many beneficial advances for engineering design, improving structure features when comparing with traditional metallic alloys. Although these alloys has been used for centuries, in many applications composite materials are substituting them partially or completely. The use of this kind of materials has produced a great impact in several areas of engineering, as transportation (aeronautics, aerospace, naval, railroad, automobile, etc.), civil construction, sport equipments, etc. An important advantage of composite materials is the possibility of compose an unlimited number of combinations of this elements thought the wide variety of matrix and reinforcements. Each combination becomes a piece with unique characteristics, able to attend specific requirements in a project. In this study, a carbon fiber plate, material used as component of the fuselage in aircrafts, is used in the experimental tests for developing a damage identification and locating method that is able to be used during the flight. The method is based on Lamb waves and it is a non-destructive evaluation (NDE). The tests were done in different conditions for a temperature range from de -45oC to 105oC. The sensing and actuation were based on piezoelectric materials (PZT), which are a versatile smart material indicated to this work since it can be utilized either as actuator or as sensor and because is very efficient at high frequencies. The process automation is realized through the application of Artificial Neural Network, since this technique has optimum robustness and capability of generalization, which are important characteristics to achieve the objective