Optimality of reinforced concrete coupled shear walls using machine learning

Abstract

A coupled shear wall is a unified system consisting by connecting two individual shear walls with a connecting beam (coupling beam). The coupling beam plays an important role in the lateral load resistance of the coupled shear wall structure. This paper addresses the innovative approach to optimising coupling beam dimensions by introducing data in machine learning. The data are collected through ETABS modelling of encompassing buildings of varying heights, i.e., 15, 20, 25 and 30 stories, with and without shear walls; coupled shear walls with coupling beams of different lengths, i.e., 1, 1.5, and 2 m, and different depths, i.e., 1.5, 1.25, 1, 0.8, and 0.75 m which are analysed by keeping the end-to-end distance of both the shear wall and the shear wall with coupled beam to make it economical. The parameters considered include displacement, drift, reinforcement quantity, and concrete volume collected through ETABS. A total of 68 models were analysed. Therefore, in all of the stories except for the 30-storey, the shear wall with a coupling beam dimension, length of 2 m and depth of 1.25 m is the best model and in the case of 30-storey optimised model changes, the coupling beam with a length of 1.5 m and depth of 1.25 m performs best. On increasing stories, it can be deduced that the coupled shear wall performs much better. Furthermore, the machine learning-trained model will provide the optimum dimension of the coupling beam if the storey height is provided.