Description
Many historical centres in Europe, including the city centre of Zagreb, feature masonry building aggregates, which developed as the building layout of the city was densified. Adjacent buildings in aggregates often share a structural wall and the building that was constructed second connects to the wall of the first building through interlocking stones or a dry joint. When compared to the response of single free standing buildings, several factors make the seismic response of masonry building aggregates more complex: It is not unusual for adjacent units to have different material properties, area and distribution of openings, roof and floor heights and orientations, and construction details. All these factors can result in an out-of-phase behaviour of adjacent units and potential separation and pounding of the units. The interface behaviour which needs to account for this interaction adds another level of uncertainty to the already complex behaviour.
It is common to model the adjacent units as isolated or fully connected, which can result in conservative approximations regarding the PGA at failure, but ignores the complex response stemming from adjacent unit interaction. To account for this, a new nD interface material model was developed and implemented into OpenSEES software.
To study the effect of interface modelling on the seismic response of a building aggregate, a case study aggregate was modelled using the Equivalent Frame Approach. The model was subjected to a non-linear time history analysis using different modelling approaches regarding the interface: fully connected units, isolated units, 1D non-linear interface, and a newly developed nD non-linear interface material model. Uncertainty of the masonry and interface parameters is taken into account to give a stochastic response. The force-displacement response of the aggregate and failure mechanisms are evaluated with regards to the interface model. The results show that the PGA leading to failure is the least sensitive of the investigated response parameters. A simplification in the representation of the interface behaviour can, however, lead to a different failure mode and failure location in the building. This becomes especially relevant when the analysis serves as input for the design of retrofit interventions.
Keywords | historical masonry, masonry aggregates, equivalent frame models, non-linear connections |
---|---|
DOI | https://doi.org/10.5592/CO/1CroCEE.2021.134 |