| Popis: |
To reduce the significant uncertainties associated with finite element modeling of masonry structures, one of the most efficient and novel approaches is the system identification using ambient vibrations and the use of modal parameters for updating of the FEM models. This paper presents system identification of stone masonry arch bridges located in the northeastern part of Turkey. The collected signals were preprocessed for system identification, and then modal parameters of the structures were identified by processing the data separately in the frequency and time domains, using Enhanced Frequency Domain Decomposition (EFDD), Auto-Regressive Model (AR Model) and Multiple Signal Classification (MUSIC) methods. The results were presented and compared for three of the representative bridges, consisting of three, four and five arches, respectively. They indicate that AR models and MUSIC method are able to capture the modal frequencies estimated by the reliable EFDD method; however, the results of MUSIC method are subject to the user’s methods, and this approach has not yet reached the required robustness for practical structural engineering applications. |
