Monitoring strategies for the analysis and preservation of architectural heritage: applications to the monumental complex of Cathedral Square in Pisa

In recent decades, structural monitoring has become a central support for the preservation of historic and artistic heritage, providing quantitative evidence to inform maintenance strategies, assess safety over time, and interpret complex behaviour under environmental and operational actions. However, heritage assets demand tailored approaches that explicitly account for conservation constraints, limited accessibility, the need for minimally invasive solutions, geometric complexity, and substantial uncertainty in both material properties and boundary conditions. This thesis presents a critical review of state-of-the-art monitoring strategies for the preservation of cultural heritage (CH) highlighting four recurrent issues. First, at an operational level, there is a lack of monitoring guidelines that define a structured procedure for the design and operation of monitoring programmes and provide specific indications for historic architecture. Second, significant challenges arise in modelling historic constructions, due to epistemic uncertainties in material properties, geometry, boundary conditions and modelling assumptions. Within this broader set of modelling issues, the role of soil-structure interaction (SSI) deserves particular attention: although often neglected in the literature, soil-induced effects may significantly influence the static and dynamic response of structures and lead to misleading interpretations if not properly assessed. For this reason, SSI is specifically investigated in this thesis, with reference to its relevance to the selected case study. Third, the integration of innovative (often non-contact) techniques remains limited in the CH context, as these methods have largely been developed for other application and have therefore not been systematically validated under the challenging small-displacement conditions typical of CH. Finally, difficulties persist in the management of heterogeneous monitoring data sources, together with a consequent lack of standardised tools for data visualisation and interpretation. To address these issues, an operational workflow is proposed, structured into phases that organise the monitoring project from preliminary investigations to data interpretation and visualisation. The workflow is conceived within a knowledge-acquisition perspective, in which monitoring and modelling are treated as mutually dependent components that are iteratively refined. The proposed framework is then applied to the monuments of Cathedral Square in Pisa, with particular emphasis on the Leaning Tower and its dynamic behaviour. Following an initial knowledge phase, encompassing historical and material analyses, past monitoring activities, and remaining open issues, detailed Finite Element Models of the Tower and the Baptistery are developed and embedded within a probabilistic framework. Key masonry and soil parameters are modelled as random variables. Uncertainty Quantification is then combined with variance-based Global Sensitivity Analysis to quantify how uncertainty propagates to the modal features and to identify the modal indicators that are most informative for monitoring purposes. Computational feasibility is ensured through surrogate modelling based on generalised Polynomial Chaos Expansion, enabling extensive sampling at reduced computational cost. Two years of continuous accelerometer recordings from the Tower are analysed using automated Operational Modal Analysis to track natural frequencies, damping ratios, and mode shapes, and to characterise their temporal variability. The dependence of the identified frequencies on environmental and operational variables is investigated, and alternative statistical strategies are evaluated to mitigate environmentally driven variability and reduce the risk of misinterpreting such variations as damage-related change. The resulting modal information is then used to update the numerical model within a Bayesian framework, assimilating monitoring evidence to infer masonry and soil parameters and to refine key modelling assumptions. In parallel, complementary monitoring technique, typically adopted in other field, are assessed for their potential in CH applications, including optical prisms surveyed by total station, ground-based interferometric radar, and satellite interferometry. Their capabilities and limitations are discussed, with practical guidance provided for cross-validation and integrated use alongside traditional measurements. The thesis concludes with the development of an integrated digital environment for long-term data consultation. A web dashboard is coupled with a three-dimensional model of the monuments, in which sensors are represented at their actual locations and linked to metadata and technical documentation. The platform provides interactive tools for visualising and exploring historical and newly acquired data, and for supporting preliminary comparisons and processing across heterogeneous monitoring sources. The original contribution of the thesis lies in combining an operational monitoring framework, uncertainty-aware numerical modelling, multi-technique validation and integration strategies, and a purpose-built digital platform for the long-term management and interpretation of heterogeneous monitoring data.