In this experimental campaign, concrete‐to‐concrete interfaces were examined under cyclic actions, crossed by postinstalled reinforcing bars or industrial steel screws. In addition, the effect of the embedment length and the roughness was investigated. It was observed that different failure modes occur, depending on the different connector types and embedment depths. The positive effect of roughness on the interface resistance was confirmed, but the results indicate that, for very high roughness values, the resistance may be smaller than for moderate values. It was also noticed that the behavior of interfaces crossed by mechanical anchors was less stiff in the ascending branch and exhibited a less steep falling branch. The extension of the project on testing thin overlays showed that the resistance was significantly affected in case of cyclic loading, suggesting that the use of any type of anchor in seismic conditions needs to be supported by adequate experimental evidence.

This study focuses on long concrete interfaces tested under cyclic actions, fastened with post-installed industrial steel screws. The overall behavior and the effect of roughness were investigated in three long interfaces, representative of connections between, e.g., a slab and a wall, a beam and a wall, etc. The results were compared with those of short interfaces tested by the authors in previous campaigns. It was observed that rough long interfaces activate their maximum resistance at small values of imposed shear slip. When roughness was reduced, the maximum resistance was also reduced, the corresponding shear slip was increased, and the overall behavior was stable. For large values of the shear slip, imposed at one end of the interface, the shear slips along it tended to be uniform, both in short and long interfaces. The limited embedment length of the screws led to their pronounced pullout. Finally, the asymmetry of resistance between the two loading directions that was observed in short interfaces was alleviated in the long ones, where also the scatter of the results was limited among duplicate specimens.