Advanced numerical models for the analysis of masonry cross vaults: A case-study in Italy

• A masonry cross-vault typology subjected to severe deterioration is analyzed.
• A full non-commercial 3D Finite Element (limit analysis and incremental non-linear) is used.
• Comparison with simplified approaches (straight and non-straight arch decomposition).
• Sensitivity analysis taking into account the role played by backfill.
• Numerical evaluation of the strengthening effect induced by FRP at the extrados.

Aim of the present paper is the analysis of a series of existing masonry cross vaults exhibiting meaningful structural deterioration and diffused crack patterns, by means of an advanced non-linear and limit analysis software.The approach utilized is a non-standard and non-commercial one and bases both for the non-linear and limit analysis procedure on a FE discretization of the domain by means of rigid infinitely resistant wedges, where all the non-linearity is concentrated on interfaces between adjoining elements.When dealing with the non-linear code, a sequential quadratic programming scheme is used at each iteration in order to deal with the deterioration of mechanical properties of interfaces, provided that the actual non-linear behavior is approximated by means of a linear piecewise constant function.Several numerical simulations are performed varying constraint conditions, material properties, infill modeling and presence of FRP strips as reinforcement devices, comparing and discussing in detail the results obtained.From simulations results, it is found that the approach commonly used in practice to study cross vaults by means of the assemblage of single arches is not always reliable, providing failure loads and mechanism quite different from the real ones. Furthermore, similarly to what occurs for masonry arch bridges, it is found that the role played by the infill is crucial and that, depending on the actual mechanical properties of the infill, both the failure mechanisms and the collapse load may vary significantly.