The natural period of vibration is one of the important dynamic characteristics that influence the seismic response of a building and the determination of seismic demands. However, its estimation in the numerical modelling of existing buildings is subject to uncertainties, particularly regarding the determination of building stiffness. Ambient vibration measurements provide an estimate of the dynamic properties of buildings at low vibration amplitudes and can therefore be used to verify numerical modelling assumptions. In this master’s thesis, three existing buildings included in the seismic stress test of the University of Ljubljana building stock were analyzed. Natural periods and mode shapes were determined from ambient vibration measurements and compared with the results of modal analyses of numerical models developed in ETABS. The study focused primarily on assumptions regarding the stiffness of structural elements and the consideration of masonry infills. The results showed that, for the NTF1 and NTF2 buildings, models with reduced stiffness due to assumed section cracking resulted in longer natural periods and larger discrepancies with the measurements, whereas the agreement improved for models without stiffness reduction. Accounting for masonry infills further improved the agreement. For the NTF2 building, the discrepancies in the first three natural periods decreased from approximately 66 %, 62 % and 57 % to 20 %, 19 % and 7 %, respectively. For the BF building, changing the stiffness assumption did not improve the agreement for all mode shapes. The results are limited by uncertainties related to the available design documentation, material properties, and simplifications in the numerical models. Further research should use nonlinear analyses to evaluate the effects of cracking and masonry infills on seismic response and to assess the relevance of ambient vibration measurements for evaluating seismic performance and designing the strengthening of existing buildings.
|
