PROPOSAL FOR A TIME-DEPENDENT DYNAMIC IDENTIFICATION ALGORITHM FOR STRUCTURAL HEALTH MONITORING

This paper describes the design, test and validation processes of a dynamic identification algorithm aimed at the time-dependent assessment of modern structures and heritage buildings for civil and seismic engineering purposes. Full validation of the algorithm is performed through analysis of numerically simulated data from an idealized masonry tower. Making use of output-only vibration measurements, the non-parametric algorithm can generate dynamic features results as time-dependent functions for the complete observation period. The algorithm can work in the presence of different dynamic loads and non-linear structural behaviours, close spectral frequency components and noise-contaminated data. Time-dependent structural dynamic parameters that can be computed are modal frequencies, modal displacements, modal curvatures, and higher derivatives of mode shapes. The proposed algorithm aims to be used as the core estimator of time-dependent identification methods devoted to the health monitoring of structures and infrastructures, being suitable for a multitude of tasks ranging from the simple operational modal analysis (in pre and post-event condition) to the complex online assessment of the structural response during seismic events for rapid damage identification.