Modelling and effects of seismic improvement interventions on church roofs

The effectiveness of interventions for seismic improvement is a fundamental issue for the preservation of heritage churches, which are especially prone to earthquake damage. Consequently, a significant research effort is currently devoted to analyze observed damage cases for structures that had undergone different kinds of intervention before the earthquake, in order to confirm and optimize those with positive effect and rule out operations that may evolve into critical situations. To this purpose, the development of realistic but manageable numerical models must face issues related to the presence of different materials, complex geometries, partial and often uncertain constraint conditions, exceptional levels of action, and the resulting nonlinearities. This work focuses on repair and improvement interventions carried out on the roof of a church, specifically San Martino dei Gualdesi, proposed by the ReLUIS Consortium as case study. The church was seriously damaged by the 2016 Central Italy earthquake but had undergone damage and repair also in previous earthquakes. In a first set of interventions the original roof had been repaired by highly increasing its mass and stiffness, according to the technique adopted at the time. Subsequently, an improvement was performed by applying a lighter roof structure, which was the configuration that underwent the 2016 earthquake. A finite element model of the whole church has been developed and static and modal analyses have been performed to characterize its behaviour, before focusing on the interaction between the church and the bell tower, attached to the church and connected to the roof structure. The bell tower was highly damaged during the event of 2016. A separate model of the tower was then adopted, after calibrating specific boundary conditions that could represent satisfactorily the continuity with the main church body and the different interconnection with the roof structure in the two cases of massive and lighter roof. Linear and nonlinear structural analyses were performed on such bell tower model. Nonlinear dynamic analyses were carried out considering three-directional motion for the October 2016 event, which was most damaging for the church, with records available from a recording station in the area. Damage development and cumulation in time could be followed for the two roof configurations. The results are discussed also in comparison with other cases that had a similar history of interventions.