Bond and Splitting: A Vexing Question

Splitting does always occur in some way prior to bond failure, in the form of either partial splitting (quite often undetected) or full splitting, the latter beng the subject of several recent papers, owing to the importance of cover splitting in R/C elements. Starting from the test results on fully-split specimens (like those by the authors on special specimens having a fabricated crack) it is possible to formulate suitable bond stress-confinement stress relationships. These models can be introduced into the limit-analysis models developed lately for the description of partial splitting up to the onset of full splitting and bar pull-out in short anchorages. In this way, a linkage between the bar-concrete pressure (studied here thoughjt a limit-analysis elastic-cohesive model) and the bond stress is estabhlished, in order to evaluate the ultimate bond capacity and to investigate the transition from a splitting-type failure to a pull-out failure. At the same time, such important topics as concrete tensile strength and fracture energy, crack cohesion and localization, concrete cover and bar diameter, fiber content and external pressure can be incorporated into the model. A set of diagrams showing the bond capacity and crack number/opening/penetration versus concrete cover is presented, and the design implications of both the theoretical and experiments results are discussed.