This research focuses on improving our understanding of the behaviour of large, lightly-reinforced concrete beams and one-way slabs subjected to shear. Empirically-based shear design methods, particularly those in the widely-used American Concrete Institute design code for concrete structures (ACI-318) do not accurately predict the behaviour of these important structural elements, and may produce unsafe designs in certain situations. Furthermore, the research community has not reached consensus on the exact mechanisms of shear transfer in reinforced concrete. This has slowed the replacement of empirically-based methods with rational methods based on modern theories of the shear behaviour of reinforced concrete. Shear failures in reinforced concrete are brittle and sudden, and typically occur with little or no warning. Furthermore, they are difficult to predict due to complex failure mechanisms. It is critical, therefore, that shear design methods for reinforced concrete be accurate, rational and theoretically sound.

An extensive experimental program consisting of load-testing thirty-seven large-scale reinforced concrete beams and slabs has been performed. The results conclusively show that the ACI shear design method can produce dangerously unsafe designs for thick concrete flexural elements constructed without transverse reinforcement. However, safe predictions of the failure loads of small-scale elements are produced. It is shown that the ACI design method does not account for the size-effect in shear, in which the shear stress causing failure decreases as the beam depth increases.

Detailed measurements of flexural and shear stresses in the experimental specimens indicated that aggregate interlock is the primary mechanism of shear transfer in slender, lightly-reinforced concrete beams. It is also shown that the size-effect can be explained by reduced aggregate interlock capacity in members with widely spaced cracks.

Digital three-dimensional topographical maps of the surfaces of failure shear cracks were constructed by scanning the surfaces with a laser profilometer. It was shown that concrete made with larger aggregate produced rougher cracks with a higher aggregate interlock capacity. The shear strength of reinforced concrete is therefore directly related to the roughness of failure shear cracks, and by extension the aggregate size, since larger aggregates produce cracks with larger asperities with improved aggregate-interlock capacity. Acoustic-emission monitoring techniques were employed to characterize fracturing in large concrete beams.

Extensive studies on the ACI 318-05 requirements for crack control steel show that they do not adequately prevent the formation of wide cracks, as they do not require a minimum bar diameter for crack control reinforcement. It is shown that the ACI 318-05 requirements for crack control steel were based partly on questionable interpretations of published experimental studies on crack widths in large beams.

Various methods to eliminate the size effect in shear are explored, including the use of stirrups or longitudinal reinforcement distributed over the beam height. Beam/slab width is shown to have no effect on failure shear stress. It is concluded that the ACI shear design method should be replaced with a rational, theoretically-sound shear design method. Modifications to Canadian shear design methods are recommended.