Design of Beam using Artificial Neural Network based Approach

Abstract

Recent developments in artificial neural network (ANN) have opened up new possibilities in the field of structural engineering. This paper demonstrates the applicability of ANN for the design of beams subjected to moment and shear. An attempt has been made to capture the mapping between the design variables using ANN. There is no direct method for design of beams. A feed forward network and back propagation training algorithm has been used. For training, about 48 data sets were generated. After successful learning, the model predicted the depth of the beam, area of steel, spacing of stirrups required for new problems with accuracy satisfying all design constraints.
Keywords: Artificial neural network, beams, steel, area, depth, moment, shear