Use of Soft Computing Techniques to Predict the Bond to Reinforcing Bars of Underwater Concrete

Abstract

The current knowledge and building code recommendations for bond behavior between concrete and embedded steel reinforcement become unwarranted when casting takes place in humid or fully submerged water conditions. The main objective of this paper is to assess the effect of underwater casting on the concrete mechanical and bond properties, including the development of regression models that simplify prediction of washout loss and corresponding bond stress-slip properties. Washout loss was determined using the standard CRD C61 test method and air-pressurized tube that simulates higher hydrostatic water pressure. Particular care was placed to reuse the same washed concrete samples for determining the residual compressive strength and bond properties to embedded steel bars. Test results showed that the washout loss and hydrostatic casting depth are directly affected by the concrete composition. Hence, for example, the decrease in water-to-cement ratio and/or addition of anti-washout admixture or silica fume improved concrete cohesiveness and resistance to washout. The stiffness and confinement of steel bars normally achieved in dry conditions curtailed with underwater casting, which led to reduced ductility and ultimate bond strength at failure. The experimental bond stress-slip data are compared with the CEB-FIP Code models and design strengths specified by ACI 318-19 and European Code EC-2. Regression analysis and predictive charts are established to facilitate assessing the effect of concrete composition on washout characteristics and residual strengths following underwater casting.