Experimental tests on post-tensioned rocking concrete walls with end columns and steel dissipative devices

In many countries, the current seismic design practice for concrete structures is based on the exploitation of the ductility capacity of some structural elements, such as beams in frame structures. This approach, even if theoretically able to guarantee life safety for the occupants of buildings, has recently been criticized because it might lead to extensive damage in structural elements. Damage that in many cases cannot be repaired. Therefore, this design practice might lead to limited resilience; as highlighted by many recent earthquakes (e.g. Christchurch earthquakes) the recovery process after strong seismic events may take years and have significant social and economic negative consequences. For these reasons different structural systems and technical solutions have been developed in the literature in order to limit structural damage, i.e. base isolation, viscous dampers, etc. Some of these technologies are very effective but might be unpractical for some types of buildings, such as precast concrete structures, in particular for economic constrains. For these buildings researchers have proposed specific solutions based on post-tensioned concrete elements free to rock at their base, with easy-to-replace external dissipative devices. These systems aim at solving the limitations of ductility-based design though self-cantering capabilities and easy reparability. The present paper presents the results of an experimental campaign aimed at studying the behaviour under cyclic horizontal loads of post-tensioned rocking concrete walls with end-columns and steel dissipative devices. The structural system tested comprises a concrete wall with two bundles of post-tensioning cables, two post-tensioned concrete endcolumns (one at each end of the wall) connected by beams and hysteretic dissipaters between the wall and each column. Columns are used in order to support beams because, even if base rocking occurs, their uplift is limited given their small cross section size. The paper discusses the quasi-static cyclic behaviour of two of these systems featuring different details at the wall-base. Furthermore, it compares the experimental results with the predictions of design-oriented analytical models.