Predicting the contribution of recycled aggregate concrete to the shear capacity of beams without transverse reinforcement using artificial neural networks
Although many researchers have studied the shear and flexural behavior of recycled aggregate concrete (RAC) beams, code provisions have not been modified yet to take RAC into account. Therefore, the slow development of code provisions to govern RAC usage limited their widespread use as a construction material for concrete structures. Several factors control the shear behavior for RAC beams and make it different from conventional concrete (CC) beams, such as recycled aggregate content and properties of parent concrete. The main objective of this study is to demonstrate the efficiency of using Artificial Neural Networks (ANNs) in predicting concrete contribution in the shear capacity of RAC beams. The study presents an appropriate model that can predict the experimental value of concrete contribution in shear resistance for RAC beams without transverse reinforcement when knowing the values of 6 inputs (recycled aggregate content, shear span-depth ratio, beam width, beam depth, longitudinal tensile steel ratio, and compressive strength at 28-days) using the intelligent adaptive ANNs based on a database comprised of 231 data points collected exclusively from structural literature. It is found that the proposed ANNs model showed satisfactory results when verified against the calculated values of the concrete shear strength calculated using common-used models existed in literature and code provisions, where the maximum variation for the present ANN model was about 8%. In particular, a comprehensive parametric study was conducted and discussed in detail to investigate the effect of various key parameters on the value of the concrete shear strength and the shape of the behavior. The results demonstrated that ANNs are capable of predicting the shear strength for beams cast with RAC without transverse reinforcement. A sensitivity analysis for the predicted concrete shear strength was conducted to give a better understanding of the effect of the key parameters (inputs).