Shear capacity of reinforced concrete deep beams using genetic algorithm

Abstract

It is vital to understand the shear behaviour of reinforced concrete (RC) beams in order to avoid a catastrophic shear failure and design for ductile failure. However, due to the complexity in the shear failure mechanism and various parameters influencing the shear behaviour of RC beams, the accuracy in the determination of the shear capacity remains a challenge. In this paper, machine learning and genetic algorithm are utilized to develop an improved shear design equation for RC deep beams without stirrups. The proposed model considers the parameters influencing the shear capacity of beams including concrete compressive strength, cross-sectional dimension of the beams, aspect ratio, and internal reinforcement ratio. The prediction capability of the proposed model has been compared with that of ACI 318 and resulted in a better prediction in terms of safety, accuracy, and economic aspects.