Aplicação de tecnologias modernas de processamento de sinais ultrassônicos com foco na inspeção de estruturas de concreto

• Ano de defesa: 2022
• Autor(a) principal: Nascimento, Thaís Eloy Guimarães
• Orientador(a): Martins, Carlos Otávio Damas
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Programa de Pós-Graduação: Pós-Graduação em Ciência e Engenharia de Materiais
• Palavras-chave em Português: Ciência dos materiais; Ultrassom; Concreto; Ensaios não destrutivos; Sinal ultrassônico; Processamento de sinais
• Palavras-chave em Inglês: Non-destructive tests; Ultrasound; Ultrasound signal; Signal processing; Concrete
• Área do conhecimento CNPq: ENGENHARIAS::ENGENHARIA DE MATERIAIS E METALURGICA
• Link de acesso: http://ri.ufs.br/jspui/handle/riufs/16687

To verify the strength and integrity of concrete structures, samples are commonly extracted or molded and subjected to destructive tests. The methodology, although traditional, can barely represent the reality of the entire structure and even destabilize it. The use of Non-Destructive Tests (NDTs) grows due to the benefits they bring, such as practicality and for allowing the evaluation of the structure in real time, without harming its stability. Ultrasound, among the non-destructive techniques, has high potential in the inspection of concrete as it allows the verification of conditions of homogeneity and compactness, but the low sensitivity to characterize internal defects, due to its limited application to the analysis of the ultrasonic pulse velocity (UPV), impacts the reliability of your field application. In this way, research focused on the applicability of ultrasound combined with signal processing methods grows, because through these tools it is possible to extract relevant information about the signal in the analyzed material, such as energy and frequency spectrum, and that are contained in the preprocessed ultrasonic data. This work sought to evaluate the sensitivity of the ultrasonic signal in concrete samples with different compositions and with internal defects through signal processing using the Wavelet Transform, and also classify the samples through Machine Learning techniques. It was possible to compare the sensitivity of the ultrasonic technique when using modern computational tools and when using only the traditional method, the VPU, as well as classifying samples according to their internal characteristics.