SEQUENTIAL SENSITIVITY ANALYSES ON MASONRY MECHANICAL PARAMETERS IN THE SUBSTRUCTURE ANALYSIS OF A MONUMENTAL CONSTRUCTION

Assessing the safety of ancient structures is particularly challenging primarily due to the difficulty of thoroughly understanding construction phases and material properties in contexts where destructive testing must be avoided. The inherent uncertainties in material parameters can generally be reduced through updating procedures, which have recently gained ground in structural engineering. In this paper, a Bayesian probabilistic approach is adopted, using existing crack patterns to deduce masonry mechanical properties and evaluate the plausibility of different construction sequences. The case study is the Baptistery of San Giovanni in Pisa, whose construction phases are mostly unknown, as masonry mechanical properties. Initially, sensitivity analyses on the uncertain masonry mechanical characteristics were performed through nonlinear models simulating possible past stages of the building. In this context, response surfaces generated through the Polynomial Chaos Expansion technique were used. Subsequently, the calibration of stiffness and strength parameters leveraging Bayesian techniques based on the observed crack patterns on the dome allowed a first validation of hypotheses regarding the construction phases. This article demonstrates how Bayesian updating techniques, coupled with sensitivity analyses, can contribute to understanding ancient structures and developing awareness of their static behavior.