Crack width prediction of RC beams by artificial neural networks

Abstract

In this paper, artificial neural networks are utilised to evaluate the cracking that occurs in reinforced concrete beams while simultaneously maintaining a consistent deformation pattern. Artificial neural networks are stimulation-based computing systems that simulate biological neural networks. Here, data from several sources is used to form the ANN model and predict the cracking behaviour of RC beams. The model ANN-(A) inspects eight different parameters that comprise height, width, effective height, cover, AS1, AS2, stress, concrete compressive strength, and primary crack spacing, respectively. The findings demonstrate that artificial neural networks (ANNs) make reliable crack width predictions, which are supported by training, testing, and validation outcomes. This crack prediction will assist in predicting the crack width of the RC beams when used with similar parameters.