Code modelling of reinforced-concrete deep beams

Code provisions for shear design of deep beams have commonly developed from limited test results. There are also very few studies on the validity of shear-carrying capacities of concrete and shear reinforcement in code provisions including strut-and-tie models. The present study investigates the effect of different parameters on the shear-carrying capacities of concrete and shear reinforcement in four code provisions: ACI 318-99 and CIRIA Guide 2 for empirical formulae, and ACI 318-05 and Eurocode (EC) 2 for strut-and-tie models. A comprehensive database having a total of 448 simple and 73 continuous deep beam test results was compiled and compared with the code provisions. Different statistical parameters to measure the safety and accuracy of code provisions are also presented for various web reinforcement arrangements. Code provisions generally fail adequately to capture the effect of different parameters on the shear capacities of concrete and shear reinforcement in deep beams, indicating that shear capacities predicted from empirical code provisions and the strut-and-tie model of ACI 318-05 are more unconservative in continuous deep beams than in simple deep beams, while the strut-and-tie model of EC2 is more unconservative in simple deep beams than in continuous deep beams. In addition, a higher unconservatism appears in strut-and-tie models than in empirical formulae.