In the recent past a number of issues regarding the seismic performance of typical existing hollowcore floor systems have been raised. The most concerning of these is the vulnerability to loss of vertical support of the floor system at the end floor to seating beam connection. This vulnerability arises due to incompatibilities between the floor system and intrinsic deformations of the neighbouring seismic frames.

In a previous contribution by the authors (Jensen et al 2006), a conceptual retrofit strategy for existing hollowcore seating connections was proposed. This paper provides an experimental validation of that strategy through quasi-static cyclic testing of alternative seating connection configurations, adopting varying seating lengths.

In general, unfavourable performance was exhibited by the existing seating connections, resulting in loss of vertical support of the hollowcore unit. In contrast, when additional seating and selective weakening retrofit approaches were implemented, a higher level of seismic performance leading to collapse prevention was achieved. In conclusion, issues and uncertainties associated with the evaluation of the likely failure mechanism, as well as the definition of an appropriate retrofit intervention are discussed.