Isospectral Equivalent Frame Models for Seismic SHM of Masonry Buildings

Advanced tasks in vibration-based structural monitoring, particularly damage quantification, often rely on the use of computational models to enhance the capabilities of data-driven techniques. In this regard, it may be essential to use efficient models to cope with the continuous flux of real-time experimental data. In several fields, high-fidelity Finite Element (FE) models are the most common choice. However, their high computational demand calls for further dimensionality reduction techniques, like surrogate modelling. For masonry buildings, a natural synthetic approach is the Equivalent Frame (EF) method, which describes masonry walls as a three-dimensional assembly of deformable beam and column elements connected by rigid nodes. This simplified description drastically lowers the model dimensions and the computational burden of linear and nonlinear analyses, trading fidelity to reduce simulation time. The work addresses the intrinsic limitations stemming from simplifying EF modelling assumptions, namely the lumping of stiffnesses and masses in a few deformable elements and the presence of rigid portions, aiming at the employment of EF as a physics-based surrogate in seismic Structural Health Monitoring (SHM) applications. These assumptions can significantly impact the estimation of modal parameters, in particular natural frequencies, which are used as a proxy for structural health. This is especially relevant for monumental buildings, often characterized by thick walls, large interstory heights, and openings arranged irregularly. By leveraging a structural identification framework recently developed by the authors, the work discusses the possibility of achieving low-frequency isospectrality—i.e., a coincident subset of natural frequencies and mode shapes—between a lumped-parameter EF model and a reference distributed-parameter FE model by means of structural identification and model calibration.