Monitoramento da integridade estrutural utilizando modelos autorregressivos
Loading...
Files
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidade Federal de Juiz de Fora (UFJF)
Abstract
The World population is constantly increasing, the evolution of society as well as the daily lifestyle require the construction of large infrastructures that meet their needs, such as: bridges, viaducts, skyscrapers, among others. Strategies based on Structural Health Monitoring (SHM) make it possible to identify damage present in structures, in addition to their respective locations and intensities, through their dynamic responses. Thus, the SHM becomes an essential tool to avoid possible structural failures and increase the useful life of structures. In this work, four autoregressive models based on time series - AR, ARMA, ARX and ARMAX - are used for the processing of dynamic signals, extracting from them parameters sensitive to structural damage in damaged and undamaged scenarios of the analysed structures. Damage levels are calculated from an indicator, whose value is obtained through the Mahalanobis Distance. These models were applied to two structures tested in laboratory: a small-sized three-floor frame and a medium-sized four-floor frame; As well as two large real-world structures, which are completed and currently in use: a 17.5 m long viaduct and a bridge with a main span of 260 m. From the analysis of results, there is a relative efficiency of the four autorregressive models in detecting structural damage. It is important to highlight the AR model ability to achieve good results, registering a relatively short code execution time. The AR model is recommended in smaller orders, if there are no dinamic excitations registered and available; and/or the usage of the ARX model, whenever dinamic excitations are registered.
Description
Citation
Collections
Endorsement
Review
Supplemented By
Referenced By
Creative Commons license
Except where otherwise noted, this item's license is described as Acesso Aberto

