Predicting the Fundamental Frequency of Masonry Towers: Insights from an Experimental Database and Slenderness-Based Clustering

In this work, an extensive database comprising more than 200 historic masonry towers and encompassing a broad range of geometrical configurations and boundary conditions, has been assembled. For each tower, the experimentally determined fundamental frequency cor-responding to the first bending mode has been documented. This dataset serves as the basis for the development of a straightforward empirical formula aimed at predicting the funda-mental natural frequency of such structures. To validate the robustness and generalizability of the proposed equation, the sample has been partitioned into training and validation sub-sets, allowing for an objective assessment of the model’s predictive performance. Multiple structural and physical parameters, including height, base dimensions, and boundary condi-tions, are explored as predictors to elucidate their influence on the dynamic behaviour of masonry towers. Furthermore, to enhance predictive accuracy, dedicated models were de-veloped for specific clusters of towers exhibiting similar slenderness ratios, rather than clus-tering them according to material composition. The results highlight the significance of slen-derness and boundary conditions in governing the dynamic response and demonstrate the reliability of the empirical approach for practical applications in the assessment and preser-vation of historic masonry towers.