Portugal, alike southern Europe, has not adequately dealt with the vulnerabilities associated to seismic risk, that occur suddenly but in a sharply manner, at an unpredictable time, location and severity. Lisbon, as an early settlement of populations, has witnessed along the centuries to a significant number of large events that have been narrated and are known to us nowadays. However, given that the time distance between events surpasses that of several generations, earthquake catastrophes tend to be forgotten... just until another one strikes. Conversely, given its centuries of History, the capital has a patrimonial value in old constructions, still standing in our days, of great importance and continuous need for preservation. Some of these constructions were built with anti-seismic concerns (for instance, empirically, care was taken to set the structures on hard rock sites and in other cases specific provisions were developed for the constructions so as to improve their seismic behaviour, the obvious example being the Pombalino buildings) while other old constructions have been designed to withstand gravity loads alone, presenting a vulnerable group at seismic risk (a recognised example would be the Gaioleiros buildings).

The response of this type of structures to strong earthquakes is still an open research subject. The analysis of the existing old masonry buildings in the light of the present seismic codes would lead to the conclusion that they are highly vulnerable to severe earthquakes; nevertheless, they behaved relatively well during past earthquakes, albeit with large variability of performances. Such discrepancies demonstrate the weakness of the available methods in accurately predicting the real behaviour of these types of buildings under destructive earthquakes. By the use of modern design techniques it is possible to apply many retrofitting schemes for the desired level of strengthening, in most cases with questionable accuracy. Moreover the level of interventions is significantly limited due to functionality and architectural reasons. Furthermore, many of these buildings belong to the historical heritage or their initial structural system should be maintained or emphasized, thus limiting the level of applicable interventions.

This project addresses the seismic hazard, the vulnerability it poses to old masonry structures and possible retrofitting solutions. Masonry can be seen as a composite material for which a great variability can be encountered as a consequence of the different materials used, the different technologies applied and the several construction systems found. For this reason the masonry constructions, other than being a complex structure to analyse have many times a regional character. Within this general framework, specific objectives will be pursued. A main concern will be the development of fragility curves based on the definition of classes of constructions for Lisbon's old masonry buildings. The development of such curves is essential as input for any loss estimation model currently being developed for the site. For such aim, analytical models will be developed of the representative construction of each class based on state of the art modelling approaches appropriate for masonry structures. Non-linear behaviour of masonry will be taken into account by running both non-linear static and dynamic analyses. Given that many of these buildings have undergone several interventions over the centuries, the most common interventions will be catalogued and modelled as well. It has been commonly recognised that experimental testing on structural components are a valuable contribute to the project. Experimental testing on relevant structural elements is also proposed given the relatively scarce knowledge and data existing on the seismic behaviour of the typical masonry constructions existing in Lisbon. Wall and pier specimens will be tested under static cyclic and monotonic horizontal loading, the obtained results will enable structural characterization of the specimens, the identification of expected failure modes and the calibration of the developed analytical models. Furthermore, retrofitting schemes will be discussed with the most interesting ones being incorporated into the analytical models.