A stochastic approach for the collapse probability of historic masonry towers

A great part of the Italian built heritage is constituted of historic masonry towers. Such structures, due to their intrinsic vulnerability are prone to seismic risk and, like every historic structure, exhibit many constructive aspects which are difficult to be caught from a numerical point of view. Among the most common: the restraint conditions is not exactly clear as they are often embedded in other buildings (such as churches bell-towers); the material behavior can’t be simply identified as the masonry walls are often multi-leaf walls, made of three layers, each one with different mechanical characteristics, etc. In this paper, with the aim to include most of the sources of uncertainties, to assess the seismic risk of historic masonry towers a probabilistic approach is presented. The probabilistic framework, originally introduced by Gusella (1991), is herein discussed through an illustrative case study: the Torre Grossa of San Gimignano, Italy. In a first part of the paper some experimental results, accelerations of some level of the tower under ambient noise, are employed to identify a numerical model. Subsequently the tuned numerical model is employed to perform non-linear time history analyses which constitute the basis for the evaluation of the Collapse Probability.