Investigation of Ultimate Shear Capacity of RC Deep Beams with Opening using Artificial Neural Networks

Deep beams are structural elements loaded as beams in which a significant amount of the load is transferred to the supports by a compression strut joining the load and the reaction. According to ACI 318-08 ration of free span to height in deep beams are less than 4. The ratio of shear span (the distance between the concentrated load to support in deep beam) is less than 2. Loading transmission mechanism in these beams is loads to reaction. The model takes into account the effects of the effective depth, shear span-to-depth ratio, modulus of elasticity, ratio of the FRP, flexural reinforcement and compressive concrete strength on shear strength. It is also goal to extend an effectual and applied neural network model. So, mechanical attribute and dimensional properties of strengthening beams are chosen as input data. A developed harmony search (HS) algorithm in ANN models is trained, validated and tested by 30 deep beams with opening. Afterward ANN results are estimated with ABAQUS results and theoretical prediction computed from ACI 318-08. Comparisons between the predicted values and 30 test data showed that the developed ANN model resulted in improved statistical parameters with better accuracy than other existing equations. Performed analysis showed that the neural network model is more accurate than the guideline equations with respect to the experimental results and can be applied satisfactorily within the range of parameters covered in this study.