| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 5.02 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
O funcionamento indevido de uma estrutura civil pode causar diversos problemas e transtornos sociais se não corrigido a tempo. Nos dias atuais, é cada vez mais necessário conhecer o comportamento estrutural, uma vez que muitas obras de arte estão no final de sua vida útil estimada em projeto. Assim, surge o conceito de monitoração da integridade estrutural, com o intuito de monitorar e antecipar problemas estruturais que possam ser gerados durante a fase de serviço e auxiliar na manutenção. As técnicas de deteção de danos fazem parte do processo de monitorização e são cada vez mais comuns quando se trata de estruturas de grande porte. Podemos distinguir estas técnicas em abordagens baseadas em modelo numérico e em dados, a partir de monitorização estática ou dinâmica, ou ainda as duas em conjunto. A monitorização estática engloba a observação de grandezas de caráter estático, como os deslocamentos, as deformações ou as rotações, enquanto a monitorização dinâmica utiliza fatores como o comportamento modal e frequências de vibração. Neste trabalho será retratado o conceito de monitorização estrutural e apresentados alguns métodos de deteção de danos baseados na resposta estática da estrutura. Dentre os métodos estudados aplicou-se a regressão linear múltipla, a análise de componentes principais e uma junção destes dois, buscando a identificação de danos na ponte em estudo, a Ponde de Lanheses, através de um gráfico de controle multivariado. O uso destes métodos permitiu também fazer uma análise da eficiência do sistema de monitorização instalado na ponte na deteção de danos.
The malfunction of a civil structure can cause various problems and social disruptions if not corrected in time. Nowadays, it is increasingly necessary to know the behavior of the structures, since many bridges and viaducts are at the end of their lifetime estimated in design. Thereby, the concept of structural health monitoring appears in order to anticipate structural problems that can be generated during the service life. Damage detection techniques are part of the monitoring process and are becoming more common when it comes to large-sized structures. We can distinguish these techniques in approaches based on numerical model and data, from static or dynamic monitoring, or both together. Static monitoring includes, among others, displacements, deformations and rotations, while dynamic monitoring uses factors such as modal behavior and vibration frequencies. In this work the concept of structural monitoring will be denoted and some methods of damage detection based on the static response of the structure will be presented. Among the studied methods, multiple linear regression, principal component analysis and a combination of these two were used to identify the damage in the studied bridge. The use of these methods has shown that the monitoring system installed on the bridge is ineffective when the damage is too far from the sensors since it is only possible to identify the damage located close to the sensors.
The malfunction of a civil structure can cause various problems and social disruptions if not corrected in time. Nowadays, it is increasingly necessary to know the behavior of the structures, since many bridges and viaducts are at the end of their lifetime estimated in design. Thereby, the concept of structural health monitoring appears in order to anticipate structural problems that can be generated during the service life. Damage detection techniques are part of the monitoring process and are becoming more common when it comes to large-sized structures. We can distinguish these techniques in approaches based on numerical model and data, from static or dynamic monitoring, or both together. Static monitoring includes, among others, displacements, deformations and rotations, while dynamic monitoring uses factors such as modal behavior and vibration frequencies. In this work the concept of structural monitoring will be denoted and some methods of damage detection based on the static response of the structure will be presented. Among the studied methods, multiple linear regression, principal component analysis and a combination of these two were used to identify the damage in the studied bridge. The use of these methods has shown that the monitoring system installed on the bridge is ineffective when the damage is too far from the sensors since it is only possible to identify the damage located close to the sensors.
Description
Keywords
Deteção de danos Resposta estática Regressão linear múltipla Análise de componentes principais Damage detection Static monitoring Multiple linear regression Principal component analysis