Abstract

Effective wall-to-floor connections are crucial components of historical buildings to avoid dangerous mechanisms under seismic actions. Existing buildings often present poor friction-based links between timber floor and masonry wall and are not able to ensure the so called “box behavior”, necessary for the correct distribution of seismic forces. Nonlinear static analysis is one of the most common tools for the seismic assessment of unreinforced masonry buildings considering advanced nonlinear materials description and allowing for different approaches. The selection of a proper control node, for the definition of the pushover curve, is fundamental and sometimes controversial. Moreover, connections are modelled as simply fixed or absent at all. Dynamic nonlinear analysis seems preferable even suffering from a higher computational effort. On the bases of previous experimental campaign developed at the University of Minho, the pull out behavior of a strengthened and unstrengthened masonry-to-timber connection was simulated numerically using OpenSees software. The connection model considers strength degradation and pinching, in agreement with the experimental behavior, and is validated from the energetic point of view, suitable for being included in a global finite element model to study the influence of the hysteretic energy dissipated within the connections on the overall seismic response. This paper describes the calibration process and the application of the connection model into a unreinforced masonry prototype using nonlinear dynamic analysis under real seismic inputs. Both strengthened and unstrengthened configurations are implemented and results compared. The selected model is part of the blind prediction competition organised within the SERA-AIMS project involving the shaking table test of a half-scaled aggregate.