A Deep Learning Based Approach for Response Prediction of Beam-like Structures

In this study, the methodology and results of ambient vibration-based investigations of the historical Tash Mosque in Kosovo and a 3-story historical building in Bulgaria are presented. The investigations include full-scale in situ testing of both structures due to ambient vibrations induced by micro-seismic, wind, traffic, and other human activities. To this aim, Ranger seismometers and Kinemetric products were used. Measurements were performed in both horizontal directions in several points along the structures' height utilizing a high-speed data acquisition device. All recorded data have been analyzed and processed by the software developed at IZIIS, and then the processed data were used as input for modal analysis. The basic assumption is that the excitation can be considered as a stationary random process to have a relatively flat spectrum. The paper clearly describes the procedure used for investigations and presents the dynamic properties of the whole structures. The investigated structures are both historical buildings and defined as architectural heritage and the outcome of this study including the natural vibration frequencies and mode shapes) can be very benefi- cial for the verification stage of the analytical/numerical models for future retro- fitting/rehabilitation schemes.