Programa computacional com interface gráfica para identificação estocástica de parâmetros modais de estruturas civis: aplicação em pontes e torres de linha de transmissão

Abstract

Interest on the dynamic behavior of latticed structures has been increased in the latest decades in Brazil since some accidents with total collapse of some structures, due to excessive vibration, have been occurred in its different regions. In the Amazon region, for instance, where these latticed structures overcome great obstacles such as the rainforest and large rives crossings, cases of collapse of latticed structures have been also reported. This master thesis focuses on the study of the modal behavior of latticed steel structures subjected to ambient vibrations caused by unknown excitation forces. Two study cases are presented: in the first, the modal behavior of a steel overhead transmission line tower is investigated; in the second case, the modal behavior and comfort level of a steel bridge is studied. The study of this last case aims at investigating the level of comfort of the bridge under ambient vibration, caused by wind gusts and vehicle traffic, according to the Brazilian standard NBR 8800 (1986). In both study cases, experimental and theoretical dynamic analyzes were performed to extract the modal parameters of theses structures. In the experimental stage, both structures were monitored using a set of low frequency accelerometers as well as a data acquisition system suitable for vibration tests of civil engineering structures. Since it is very difficult to measure the magnitude of ambient excitation forces, two stochastic subspace identification methods known as SSI-DATA and SSI-COV were employed to extract the modal parameters of the analyzed structures from response data collected in the vibration tests. Among all activities performed in this stage, a Graphical User Interface (GUI) toolbox in Matlab® platform was developed. In this toolbox, the above mentioned stochastic methods were implemented to extract modal parameters of civil engineering structures only from the response data collected in vibration tests. This toolbox consists of three modules: the first is used to process and treat the signals collected in the vibrations tests; the second is designed for allowing user to add the information related to the position and orientation of the accelerometers to the data files; the last module is used to extract the modal parameters from the response data files preprocessed in the first two modules. In the theoretical stage, numerical models were created to simulate de dynamical behavior of the structures under study. By comparing the results from both stages of analysis, it is observed that the experimental and theoretical results are in good agreement for the first modes of vibration. The experimental results showed that the approach developed for output-only modal analysis of civil engineering structures using stochastic state-space based methods was very efficient for identification of the modal parameters of the analyzed structures.